There is a conversation about treaty relief that founders with cross-border operations consistently avoid, and it is not the one about rates. It is the one about whether they are entitled to the treaty at all.
Everyone wants to talk about the lower withholding percentage. Almost nobody wants to talk about substance, beneficial ownership, and what happens when the tax officer asks why the entity receiving the payment exists.
The comfortable version of the conversation
It goes like this. The founder has an entity in a jurisdiction with a favourable treaty. Payments flow from India to that entity. The treaty rate is lower than the domestic rate. Everyone is pleased, the number in the model looks better, and the discussion ends.
What was skipped: does the receiving entity actually have a claim to that treaty? Is it a resident of that country for treaty purposes, does it have genuine substance there, and is it the beneficial owner of the income, or is it a conduit passing value on to someone else?
Why those questions now decide the outcome
Treaty access has tightened considerably. The direction of travel across almost every jurisdiction is the same: benefits are for arrangements with commercial purpose, not for structures whose principal purpose is obtaining the benefit itself.
In practical terms, the questions I expect to be asked about an intermediate entity are:
- Does it have people, premises and decisions of its own, or only a registered address?
- Do its directors actually meet there and actually decide anything?
- Does it bear risk and hold assets, or does income arrive and leave largely unchanged?
- Is there a commercial reason for it beyond the tax rate?
- Does it have a tax residency certificate, and can it support what that certificate implies?
A structure that cannot answer these confidently is not a tax-efficient structure. It is an exposure that has not been assessed yet.
Where founders get caught
The pattern I see most is a structure inherited rather than designed. It came with an early investor’s preference, or a previous adviser’s template, or it was copied from a company at a different stage. Nobody currently in the business can explain why it is shaped that way, which is a difficult position from which to defend it.
The second pattern is substance that decayed. The entity had a real function once. The team moved, the operations consolidated somewhere else, and the entity remained on the diagram receiving income while doing progressively less. Substance is not a one-time test; it is a description of how things currently work.
The third is characterisation. Whether a payment is a royalty, a fee for technical services, business profits or something else determines which article applies and often whether tax is due in India at all. This is decided by what is actually provided, not by the invoice description. Software, data access and cloud arrangements are where this goes wrong most frequently, because the commercial reality and the contract language have often drifted apart.
The documentation nobody wants to prepare
Treaty relief in India is procedural as well as substantive. The residency certificate, the prescribed declaration, the certification supporting the remittance — these are prepared before the money moves. There is no retrospective fix that is as clean as doing it in order.
This is unglamorous work and it is where the relief is actually won or lost. I have seen entirely legitimate treaty claims fail on paperwork, and I have seen aggressive ones survive because the file was immaculate and the commercial story was consistent.
How I would have the conversation
Start from the business, not the structure. What does each entity in the group actually do? Who makes its decisions and where do they sit? What value does it add to the flow of income? Then ask whether the structure on paper matches those answers.
Where it does, document it properly and claim the relief with confidence. Where it does not, you have two honest choices: give the entity real substance, or collapse it. The third option — leaving it in place and hoping the question is not asked — is the one founders default to, and it is the one that turns a manageable restructuring into a multi-year assessment.
The cost of avoiding it
Treaty positions are reviewed years after the fact, when the business is larger and the stakes are higher, and often during diligence when there is a deal on the table and no time to fix anything. That is the worst possible moment to discover that the structure cannot answer basic questions about itself.
Have the conversation while it is still a planning exercise. It is considerably cheaper than having it as a disclosure.
If your group has intermediate entities and nobody has stress-tested the treaty position recently, AeTx does exactly that review, including the substance questions most advisers leave until they are asked by someone else.