The Madras HC transfer-pricing CA and the Korean subsidiary audit

His family is old Chennai—his grandfather was a judge on the Madras High Court bench in the 1970s, and Surya's library has inherited a complete run of Tamil Nadu Law Reports from 1962 onwards, alongside a working collection of CBDT circulars, advance ruling compendiums, and ten years of marked-up court judgments in customs and transfer pricing. He lives with his wife Anjali, a radiologist at Apollo Hospitals, and two children in a 1970s bungalow in Alwarpet with a teak-lined study that smells permanently of leather bindings and South Indian coffee. He is not a man who makes quick decisions. He makes written decisions, in triplicate, with margin notes.

His monthly inflows are straightforward: ₹1,80,000 per month from his accounting practice (nine clients, two permanent and seven seasonal — the Hyundai subsidiaries in particular spike around March 31); another ₹45,000 from speaking fees at the ICAI Chennai chapter; and Anjali's hospital salary of ₹1,10,000. It is more than enough. What is unusual is what happened in January 2026 when a single email from the DCIT's office triggered two parallel audits, and Surya discovered that the rule governing the component's value in income tax and the rule governing it in customs law were operating in opposite directions.

🗓️ The regulatory labyrinth

Transfer pricing in India is deceptively simple on paper and crushingly complex in practice. When a multinational company like Hyundai Motor's Korean parent sells a steering-column subassembly to Hyundai Motor India Limited—Surya's client—the price it charges is not a market price. It is an internal price, set by the company's finance department, and it must be documented to prove it would have been charged in an arm's length transaction between independent parties. This is the Transfer Pricing regulation. It lives in Rule 10D of the Income Tax Rules, 1962, and it requires a "comparable uncontrolled price" — usually gathered from a benchmarking study that compares the transaction to similar transactions in the market.

But when the same steering-column subassembly arrives at the Chennai Port, its value is not a TP benchmark. It is subject to the Customs Valuation Rules, 2007, which ask: what would this component cost if it arrived as a spot purchase from an independent supplier, on the same day, to the same importer? The Customs authority has a different definition of "transaction value" — it is literally the price on the invoice, adjusted for freight, insurance, and royalties, but it must be verified against independent transactions in the same month.

The income-tax rule asks: would an independent party charge this price to Hyundai? The customs rule asks: what price did independent suppliers charge for the same component in the same port that month? They are not the same question. For eighteen years, Surya had navigated the distinction by writing two separate TP documents — one for income tax and one for customs — with sufficient overlap that he could defend them in parallel if challenged. It had worked. In January 2026, it stopped working.

1. ⚖️

   Dec 2025 — DCIT TP audit notice

   The Directorate of Income Tax (Transfer Pricing) selected Hyundai Motor India for a Section 92(1) TP audit covering FY 2024-25. The component in question: Korean-sourced steering-column subassembly, invoiced at ₹18,900 per unit.

2. 📨

   Jan 2026 — Customs simultaneous demand

   Chennai Port Authority's customs officer issued a SCN (Show Cause Notice) the same month, claiming the component's customs valuation was overstated for ITC purposes, and issued a provisional assessment at ₹14,200 per unit — a 25% discount.

3. 🛑

   Feb 2026 — Conflicting benchmarks

   DCIT's valuation expert cited Hyundai's own comparable uncontrolled price from a Japanese supplier (₹19,500). Customs cited a Chinese competitor's average (₹13,800). The two documents contradicted each other — one proved the price was too high for TP, the other proved it was too low for customs duty.

4. ₹

   May 2026 — Consolidated resolution

   Surya's agent-assisted dual-track response: separate TP benchmarking addendum addressing DCIT comparables; customs duty rebuttal with port transaction data and exchange-rate parity logic. Dual appeals filed 10 days apart. CESTAT admission within 3 weeks.

What very nearly happened was a year of fire-fighting. The DCIT would demand that Hyundai prove the transfer price (₹18,900) was arm's length. If Surya successfully proved it, the customs officer would point to the same proof and say: "Your own documentation shows this price is justified—therefore the customs valuation is also ₹18,900, not the ₹14,200 my valuation expert cited." Conversely, if Surya conceded the customs demand and repriced downward to ₹14,200, the DCIT would claim Surya had admitted the component's actual value was ₹14,200—which meant the TP price of ₹18,900 was overstated by 33%, and Hyundai owed back taxes on that overstatement, plus interest, plus penalty.

He had seen this game before. A textile exporter in Tiruppur had spent fourteen months caught in the same loop. A pharmaceutical importer in Coimbatore had conceded on one front and lost on the other. The precedent in the courts was thin because most companies simply paid.

Surya drew a timeline on his iPad Pro in red and blue pen: red for DCIT deadlines, blue for customs deadlines. The red and blue deadlines crossed at March 31 — the same date both responses were due. He circled it three times.

"நான் இரண்டு அறிஞர்களுக்கு ₹50,000 செலுத்த வேண்டும் — ஒருவர் ₹18,900 ஆக வாதிட, மற்றவர் ₹14,200 ஆக வாதிட."

— I need to pay two counsel ₹50,000 each — one to argue the price is ₹18,900, the other to argue it is ₹14,200.

He did not tell Anjali about the timeline. She knew the outlines — she had sat through enough ICAI coffee-table conversations to know what concurrent audits meant. What he did not tell her was that his fee from this one defense, if it went to Madras HC and back, could be ₹8 lakh across two years, and the contingency was that he lost on one front and won on the other, which meant the net result was no fee at all.

⚠️ What very nearly happened

The DCIT's TP audit notice arrived on January 8, 2026, on his mobile as an SMS from the Income Tax e-filing portal. The customs SCN arrived a week later, hand-carried by a customs officer to Hyundai's import section at Chennai Port. Both notices cited the same invoice line — steering-column subassembly, import order HYUNDAI-KOR-2024-0847, ₹18,900 per unit.

Surya called his in-house counsel, Deepak Srinivasan (₹28,000 per month retainer for day-to-day advice), who said: this is not day-to-day. He called his senior advocate at the Madras HC bar, V. Anantha Kumar (₹4,200 per hour, minimum three hours = ₹12,600 just to listen), who said: you need a customs counsel separately because the TP regulation and customs law operate on different precedent trees. That would be another advocate, Siddharth Iyer (₹3,800 per hour), minimum engagement ₹15,000.

So Surya was looking at three counsel, a document-drafting fee for himself (estimated ₹2 lakh), and a dual-track timeline that would mean two separate benchmarking studies, two separate defenses, and two separate appellate strategies. The DCIT would want to hear why Hyundai's price compared to Japanese and Korean OEM parent TP policies. The customs officer would want to hear why the price compared to spot purchases from Chinese and Taiwanese suppliers in the same month.

The documents contradicted. The TP benchmark said ₹18,900 was defensible with the Korean parent's comparable uncontrolled price methodology. The customs comparable said ₹14,200 was defensible with the Chinese supplier's average spot-market rate. Surya had a choice: concede one front and defend the other, or try to defend both and likely lose both because the logic was incoherent.

He printed both notices and left them on his teak desk for three days.

🌗 What changed

On a Friday in February, Anjali mentioned, lightly, that her colleague at Apollo — a radiologist named Dr. Krishnaprabhu — had mentioned that his cousin, a retired CESTAT judge, had recommended an AI agent for "the regulatory reading part" of customs cases. She said it without expecting Surya to listen. Surya, who trusts his library more than his browser, listened anyway. That evening, he installed the agent on his iPad Pro and typed, in English, a single sentence:

"I am a transfer-pricing CA in Chennai with a concurrent DCIT audit and a customs SCN, both on the same invoice. The DCIT's comparable price is ₹18,900. The customs comparable is ₹14,200. I need to prove both prices are correct in their own logic without contradicting each other. What is the legal principle that distinguishes TP documentation from customs valuation such that I can defend both prices simultaneously?"

The agent read Section 92 of the Income Tax Act and Rule 10D of the Income Tax Rules. It read the Customs Valuation Rules, 2007. It read ten years of Madras HC judgments on the subject. It returned, in ten minutes, a principle Surya had not seen articulated in his own library:

"Transfer pricing and customs valuation serve different legal purposes, and the same invoice can be defensible under both. TP asks: would an independent party charge this price on this date to this buyer, given the buyer's bargaining position and the market structure at the time? Customs asks: what did independent suppliers charge on similar transactions in that week, in that port, to different buyers? If Hyundai paid ₹18,900 to the Korean parent and that price is defensible as arm's length in its own market structure, then the TP is solid. If a Chinese supplier charges different buyers ₹13,500–₹14,500 in the same week, then the customs valuation is also solid—because customs valuation is market-price weighted, not transaction-specific. The DCIT should be asked: does your comparable-price methodology account for the bargaining power difference between Hyundai (a captive buyer) and independent Chinese suppliers (commodity sellers)? The customs officer should be asked: can you prove the Chinese supplier prices are from arm's length spot purchases in the same port, or are they third-party price surveys with unknown buyer-seller relationships? These are different questions. You can answer both."

Surya read it twice. He read it on his iPad. He printed it and read it again, with a red pen in hand, marking the logic tree. The principle was not wrong. It was a way of re-framing the problem so that the two audits were not in logical conflict—they were in parallel jurisdictions with parallel-but-different standards.

He called Anantha Kumar, his senior advocate:

"I have a principle. Transfer pricing is transaction-specific; customs valuation is market-weighted. If I can show Hyundai paid ₹18,900 and that's defensible in its own buyer-specific context, and I can show the customs data is from commodity sales with different buyer structures, I can defend both without conceding either. Can we reframe the DCIT response around buyer-bargaining-power parity?"

Anantha Kumar said: yes, this principle exists in TP jurisprudence. It is in the Delhi HC judgment in Vodafone 2012 and the Chennai bench decision in Orient Cement 2019. But it does not appear in your DCIT audit notice, and the customs officer is not trained to hear it. You will need to introduce it.

Surya spent the next ten days rewriting the DCIT response. Instead of defending the ₹18,900 price as arm's length in absolute terms, he defended it as arm's length given Hyundai's specific bargaining position as the Korean parent's captive buyer. He then prepared a separate customs rebuttal, not conceding the DCIT's logic, but instead arguing that the customs officer's ₹14,200 comparable was based on independent supplier prices for commodity sales — a different buyer-seller relationship — and therefore not comparable to Hyundai's captive-buyer scenario.

The two documents no longer contradicted. They occupied different analytical spaces.

⚖️

DCIT Transfer Pricing

₹18,900 — arm's length in TP context

DCIT asks: would an independent party in Hyundai's buyer position charge this price? Answer: yes, because Hyundai's captive-buyer status justifies a premium over commodity prices. The TP is defensible.

📦

Customs Valuation

₹13,500–₹14,500 — market-weighted comparable

Customs asks: what did suppliers charge independent buyers in this port-month? Answer: ₹13,500–₹14,500, but only for commodity sales. These are not comparable to captive-buyer pricing structures.

📋

Surya's Reframe

Both defensible — different jurisdictions

Neither auditor is wrong. They are applying different standards to different buyer-seller relationships. Surya's documentation now proves both prices are correct in their own analytical frame, without contradicting the other.

He filed the DCIT response on March 15 — fourteen days before the deadline. The customs rebuttal followed on March 25, citing the agent's principle directly: "These transactions operate under different buyer-seller relationship frameworks, and the benchmarks are therefore not contradictory in nature."

On April 3, the DCIT's examination officer acknowledged the reframing and asked for a revised benchmarking addendum. On April 8, the customs officer issued a preliminary assessment at ₹17,200 per unit — a compromise, but defensible. Neither the DCIT nor the customs officer had conceded the full argument, but both had stepped back from the hardest positions.

The agent had not written the legal argument. Surya had written it. The agent had done what no counsel could do in three hours for ₹12,600: it had read ten years of fragmented Madras HC jurisprudence, identified a principle Surya's library contained but his memory had not surfaced, and presented it as a navigable path between two impossible choices.

🧭 Why we built it

There are, by the ICAI's registry, approximately 1,200 chartered accountants specializing in transfer pricing across India. Among them, perhaps 150 work concurrently with customs disputes — not because the speciality is crowded, but because the two regulations are taught separately in CA curriculum, and most practitioners do not maintain deep jurisprudence on both. Tamil Nadu, because of its port infrastructure and auto sector, has perhaps 20 such specialists. Among those, most are in firms large enough to have a customs advocate in-house. Surya is a solo practitioner with a retainer advocate and a rolodex of expert consultants he calls on need-basis. The gap is not in his competence; it is in the time-cost of maintaining multiple specialist libraries in parallel.

The problem Surya's case illustrates is that the safe decision-making process is also the expensive one. A big-4 accounting firm would assign this problem to three specialists: one from TP, one from customs, one from international tax. They would sit together, read the same notices, and compare notes. Surya, doing this alone, had to pay for that comparative reading from outside counsel. The only way he could afford to do it was to hire advocates at ₹4,200–₹3,800 per hour and let them work in parallel, which cost ₹40,000 before he had a defensible strategy.

What he needed was not a lawyer — it was a research partner who could read the regulations fast, compare them to case law, and surface the principle that bridges them. An AI agent sitting at the desk beside him, reading the Madras HC judgment from 2019 that contained the principle, costs zero per hour. The time-to-insight went from "call counsel, wait for a callback, book a meeting, pay for three hours" to "open the agent, type the problem, read the answer in ten minutes".

The principle the agent surfaced — that TP and customs operate in different analytical frames — was not novel. It was in the courts. It was in Surya's library. What was novel was the speed at which an agent could retrieve it from a jurisprudential haystack and present it not as a fragment, but as a navigable path.

What it does

• 🔍Reads ten years of Madras HC and Apex Court judgments on TP/customs distinctions and retrieves principles the litigator may not recall under deadline pressure.
• 🗂️Cross-references DCIT TP methodology (Rule 10D) against Customs Valuation logic (Customs Act Section 14) and identifies which jurisprudential precedent applies to which regulation.
• 📞Identifies when the same invoice triggers contradictory positions in two jurisdictions and surface the legal principle that converts contradiction into parallel defensibility.

What it does not do

• 🔒Never files a response, signs a document, or commits to a legal strategy without Surya's explicit review and approval at each step.
• 💳Never submits a customs rebuttal, a TP benchmarking study, or an advance ruling application without Surya signing off in his own professional name.
• ✅Never decides which comparable to cite, which precedent to prioritize, or which auditor to push back against — it surfaces principles; Surya makes the advocacy choice.

We built it free. We will keep it free for this user — the transfer-pricing specialist in a T. Nagar office who sits beside a library of hand-marked Tamil Nadu Law Reports and trusts the regulation more than the portal — forever. Transfer pricing is not a market segment with venture funding. It is a market segment with solo practitioners, small partnerships, and firms two sizes too small to have a customs advocate and a TP advocate on the same payroll. Surya will eventually pay for premium features: a port-specific customs comparable database, a monthly CBDT circular digest in Tamil, a Madras HC judgment alert service keyed to his client list. But the core thing — reading the ten-year jurisprudence beside him when the audit notice arrives, in minutes, not in counsel fees — is, and will remain, free.

🌱 What we hope happens

Deepak Srinivasan, Surya's in-house counsel, sent us a message in April, a few days after the preliminary customs assessment landed at ₹17,200. He said Surya had not changed his manner; that was not the point. The point was that on the Tuesday morning after the assessment, as Deepak was preparing the next defense, Surya had called him and said: before you hire a third counsel, let the agent read the CESTAT precedent on valuation appeals, because the customs officer may try to push the assessment upward at the appellate stage, and I need to know what the strongest precedent is.

Which is what we hoped this would be. Not a rescue. Not a phone call to an advocate who is busy with another brief. Just a tablet on the teak desk in T. Nagar, set to English, that reads the regulation alongside Surya when the auditor's letter arrives and says: the principle that divides these two jurisdictions is in the Delhi HC judgment from 2012; the Madras bench applied it in 2019; your invoice is defensible under both rules simultaneously.

There is a specific kind of invisible time cost in Indian specialist practice that does not announce itself in rupees. Nobody took anything. The DCIT issued a routine audit notice. The customs officer issued a routine SCN. Both officers were doing their job. The gap was not in the regulation or the courts; it was in the time-cost of maintaining two specialist libraries in parallel, reading them under deadline pressure, and extracting the principle that bridges them. For practitioners whose sons told them to automate the reading — which is most of the conscientious ones — the principle was there, but invisible.

If you have a practice in concurrent taxation disputes, GST-with-customs ITC challenges, or TP audits concurrent with customs duty demands, the agent is free at gabforge.in. We have an Indian tax regulation codex with Madras HC, Delhi HC, Bombay HC, and Apex Court precedent from 1986 onwards, and the routing knows CESTAT procedure, DCIT transfer pricing methodology, and GST valuation disputes at each major port. You can set it up on a tablet in twenty minutes. We will not advertise to your clients. We will not sell their confidential audit notices. We will read the regulation with you — the regulation that has been in your library for eighteen years but arrived as a principle at exactly 11:47 PM on a Thursday when counsel's office was closed — and we will be silent.