The Foreign Subsidiary in the Group Pack: Currency, Dates, and the Details That Bite

The pattern is common in the Indian mid-market and getting more so. An Indian group sets up a Delaware entity to bill US customers, or a Singapore company to hold the region, or a UAE arm for the Gulf business. Sometimes it runs the other way: a foreign parent with Indian operating companies that need rolling up. Either way, the group pack now contains an entity whose books live in another currency, often on another calendar, usually in another system — and the consolidation, which was already the hardest part of the pack, acquires a new layer of ways to be quietly wrong.

None of this is exotic accounting. The rules for translating a foreign operation are settled and precise. The trouble is that the hand-rolled version — a tab in the consolidation workbook with a rate typed into a cell — tends to honour those rules approximately, and approximately is exactly where the errors live. What follows is the short list of details that bite, and what honest handling of each one looks like.

Three rates, not one

The first discipline is that a foreign subsidiary is not translated at “the exchange rate.” It is translated at three different rates, applied to three different parts of the accounts, for three different reasons.

The income statement translates at the average rate for the period. Revenue and costs accumulate across the month or quarter, so the average rate is the honest proxy for the rates that actually applied when the transactions happened. The balance sheet translates at the closing rate. Assets and liabilities are a snapshot at a point in time, so the rate at that point is the right one. Equity translates at historical rates. Share capital was invested on a particular date at a particular rate, and that is the rate it keeps — the parent’s investment doesn’t grow because the rupee moved last month.

Mix these up and you manufacture results. Suppose the US subsidiary earned $10,00,000 evenly across a quarter in which the rate drifted from ₹83 to ₹87, averaging ₹85. Translated at the average, that’s revenue of ₹8.5 crore. Translated lazily at the closing rate, it’s ₹8.7 crore — ₹20,00,000 of revenue that no customer ever paid and no invoice ever carried, created entirely by picking the wrong rate. The subsidiary’s performance didn’t change; the spreadsheet’s convention did. A phantom gain of that kind is worse than a visible error, because it looks like growth.

The residual belongs in equity, not in “other expenses”

Apply the three rates correctly and something inconvenient happens: the translated balance sheet no longer balances. It can’t — you’ve translated different parts of the same accounts at different rates. The difference that emerges is the cumulative translation residual, and it is not a mistake. It’s the arithmetic consequence of restating an investment held in one currency into another, and it belongs in equity, as a translation reserve, sitting apart from the trading results.

The classic hand-rolled error is to treat that residual as a nuisance and plug it wherever it fits — most often into “other expenses” or “other income” to force the balance sheet to tie. This does two kinds of damage. First, it contaminates the operating result with a number that has nothing to do with operations: the group’s reported margin now moves with the currency, and nobody reviewing the pack can explain why April’s costs jumped. Second, it hides the residual itself, which is genuinely informative — a growing translation reserve tells the board something real about currency exposure on the foreign investment. A plug buried in an expense line tells them nothing, wrongly.

The test is simple. If someone asks “why did other expenses move ₹14,00,000 this month?”, the answer should never be “that’s the balancing figure from the US translation.” If it is, the consolidation is presenting a plug as a fact.

When the period-ends don’t line up

The second detail is calendrical. The Indian group closes to an April–March year; the US subsidiary may run January–December because its local compliance does. Even where the year-ends align, the rhythm often doesn’t — the subsidiary’s bookkeeper closes on a different schedule, and the books for June arrive in the second week of August, after the group pack has gone out.

There are two ways to handle this, and only one of them is honest. The dishonest way — near-universal in hand-rolled packs — is to mix silently: consolidate June for the Indian entities with May for the US arm, or roll the subsidiary’s prior month forward, and label the result “June, consolidated” as if every line were equally fresh. The pack looks complete. It isn’t, and nobody reading it can tell.

The honest way is to flag the stale period rather than hide it. If the group pack for June carries the US entity at May, say so — on the page, next to the numbers, not in a footnote nobody reads. A reader who knows one column is a month old can weight it accordingly. A reader who doesn’t know is simply misinformed, with a precision that makes the misinformation convincing. Staleness disclosed is a limitation; staleness concealed is an error.

Inter-company balances that stop matching

Within one currency, inter-company elimination is demanding but mechanical: the parent’s receivable from the subsidiary and the subsidiary’s payable to the parent should be equal and opposite, and on consolidation they vanish. Add a currency boundary and they stop matching by default. The parent booked the loan in rupees; the subsidiary carries it in dollars; translate the dollar side at the closing rate and the two legs differ — because of the rate movement, or a settlement in transit at period-end, or an invoice one side has booked and the other hasn’t.

What matters is what the consolidation does with that difference. The tempting move is to absorb it — adjust whichever side was keyed last until the elimination nets to zero, and move on. That makes the mismatch disappear without ever establishing what it was. A rate-driven difference is expected and explainable; a difference caused by a missing entry is a bookkeeping error that will compound. Absorbing them together means never knowing which you had.

Matched-pair elimination should do the opposite: pair the two legs, eliminate the matched amount, and surface whatever remains as a visible residual with a stated cause. A residual of ₹3,40,000 labelled “closing-rate difference on inter-company loan” is a consolidation working correctly. The same amount silently netted away is a question nobody got to ask.

The data problems underneath

All of the above assumes you can get the subsidiary’s numbers into the consolidation at all, and that assumption does real work. The Indian entities keep books in Tally, in lakhs and crores against an April–March year; the US arm is in QuickBooks or Xero, in dollars and thousands, on a chart of accounts that shares nothing with the group’s beyond good intentions. Between the two sits a mapping — this account of theirs is that account of ours — that typically lives in a spreadsheet, maintained by hand, by one person, updated whenever someone remembers that the subsidiary added an account three months ago.

Every fragility in that chain surfaces as a consolidation error later. An unmapped account silently drops out of the group total. A mapping changed mid-year makes the comparatives incoherent. Rates keyed by hand from whatever source was open that day mean the same quarter translates differently depending on who ran it. None of these is an accounting failure; they’re data failures. But the board reading the pack can’t tell the difference, and shouldn’t have to.

What good looks like

The standard for a foreign subsidiary in the group pack is the same standard as for everything else in it: reproducible, inspectable, and honest about its own limits.

Concretely: the rates used each period are documented — which rate, from which source, for which purpose — rather than living in a cell. The translation is reproducible: the same entity books and the same rates produce the same group figures, every time, run by anyone. The translation residual sits in equity where it belongs, visible and trended, not smeared into operating lines. Eliminations show their work — matched pairs eliminated, residuals surfaced and explained. Stale periods are flagged on the face of the pack. And any group number drills back through the translation and the mapping to the entity, and the ledger, that produced it. That chain — group figure to entity figure to underlying books — is what makes a consolidated number answerable rather than merely presentable.

What to expect with Datavrn

Datavrn is being built to treat foreign entities as a defined part of the consolidation, not an appendix to it. Rates are recorded per period and applied by rule — average to the income statement, closing to the balance sheet, historical to equity — so the translation is reproducible rather than re-keyed. The residual goes to a translation reserve automatically, never to an expense line. Inter-company eliminations across currencies surface their mismatches as visible, explained residuals. Period alignment is explicit: if an entity’s books are a month behind, the pack says so rather than pretending otherwise. And the group figure drills back through the mapping to the entity that produced it. The judgement about the group’s exposure stays with you; the arithmetic that has to be exactly right stops depending on which cell got typed last.

Datavrn is management-reporting software for finance teams, virtual CFO practices and enterprise finance functions. Available now: for one standalone Entity, take a Trial Balance through confirmed groupings and disclosures to an Excel working draft with live links or a clean finalised workbook with fixed values. Datavrn prepares the statements; it does not file or certify them. The wider management pack is in active development. Start free with your work email — no invitation needed.