What Happens If a Branch Keeps No Separate Books in the UAE?

The branch without its own books is the taxable presence that cannot be measured: the UAE operations' revenues unseparated from the head office's, the costs unallocated (or arbitrarily allocated), the assets and liabilities unattributed, the month's profit unknown — the branch's economics dissolved in the consolidation.

The tax framework needs the measurable branch: the profit attributable to the UAE operations — the functions performed, assets used, risks assumed in the Emirates — determined arm's-length, taxed accordingly. The attribution is the accounting exercise built on the branch's books; without the books, the attribution is the assertion, and the authority's estimate fills the vacuum.

The PE dimension: the branch's books evidence the permanent-establishment position — the activity's nature, scale and duration provable from the branch's own records. The unbooked branch's PE story is told without the documents; the booked branch's with them. The unmeasurable branch is the undefended branch — the tax base, the deductions, the PE all unproved.

The carve-out constructs the branch's accounting reality: the UAE transactions identified and recorded separately — the branch's sales (the contracts performed in the UAE), the purchases, the payroll (the UAE people), the expenses incurred for the branch — each in the branch's books as they occur, not the head office's year-end allocation.

The balance sheet attributed: the branch's working capital (the receivables, the payables, the cash), the fixed assets used in the UAE, the inter-branch balances (the head-office funding, the clearing accounts) — the branch's financial position, complete and provable.

The opening established: the reconstructed opening (where the history was unbooked — the bank records, the contracts, the head-office data as the backbone), honestly labelled by evidence quality. The branch's trial balance produced — the UAE operations', monthly, complete. The carve-out is the accounting project; the monthly branch books are its permanent product.

The branch's profit attribution (verify the applicable framework): the functional analysis — the functions performed in the UAE, the assets used, the risks assumed — the branch's economic substance identified; the dealings with the head office priced as between independents — the services, the funding, the intangibles at the arm's-length price.

The head-office charges disciplined: the allocation keys documented and benefit-reflecting (not the plug), the charges recorded as the branch's real costs, the benchmarking where the framework requires. The arbitrary head-office charge is the disallowance waiting — the evidenced arm's-length charge is the deduction.

The transfer-pricing documentation: the branch's file — the functional analysis, the pricing methods, the dealings documented — the attribution's evidence pack. The booked branch with its file defends its profit; the unbooked branch defends nothing. The arm's-length discipline is the branch's tax credibility.

The branch's books double as the permanent-establishment evidence: the activity recorded — the functions, the people (the UAE headcount, the roles), the contracts (concluded in the UAE, the negotiation trail), the duration (the permanent presence, not the fleeting) — the PE's existence, nature and scale provable.

The treaty analysis applied to the evidenced facts: the applicable tax treaty's PE article (verify), the agency-PE and fixed-place questions answered on the records, the profit attributable to the PE determined through the attribution. The booked branch's PE position is the evidenced position — the analysis the authority can follow.

The unbooked branch's PE position, by contrast, is constructed by the authority from the outside: the contracts, the people, the presence — observed, not provided. The branch that evidences its own PE controls the narrative; the branch that doesn't receives the authority's version. Book the branch; evidence the PE; control the story.

Do not allocate the profit by formula at year-end. The branch's profit 'determined' as the fixed percentage of the group's profit (the formulaic attribution without the functional analysis) — the arbitrary carve-up, indefensible under the arm's-length standard. The attribution follows the functions, assets and risks — evidenced, not formulaic.

Do not leave the inter-branch dealings unpriced. The head-office services consumed free, the funding uncharged, the intangibles unlicensed — the unpriced dealings distorting the branch's profit (and the head office's). Every dealing priced arm's-length, recorded, documented — the branch's profit real because its costs are real.

Do not treat the branch as the cost centre. The branch recorded as the pure cost accumulator (all costs, no revenue attribution — the 'support office' fiction where the revenue-generating activity sits) — the profit understated by the construction, the attribution challenged on sight. The branch's economics recorded truly: the revenues it generates, the costs it incurs, the profit it earns.

The branch with its books built holds the measured presence: the UAE operations' profit determinable and defended (the attribution evidenced, the TP file complete), the deductions provable (the branch's costs recorded), the PE position evidenced (the presence provable from its own records).

The measured presence's dividend: the tax filings as the reports of the branch's books (not the group's allocation), the audit passable (the branch's records existing), the head-office relationship clean (the dealings priced, the balances agreed), the banking and licensing supported (the branch's financial position producible).

And the monthly truth: the branch's trial balance — the UAE operations seen monthly, the profit known, the attribution current. The branch ceases to be the consolidation's footnote and becomes the measured business it is — the taxable presence, properly measured, fully defended. Build the books; the measurement follows.

Nordvik Engineering AS, a Norwegian company, operates a Dubai branch executing regional projects. The branch keeps no separate books: project costs are paid from head-office accounts, branch staff salaries are processed through the Oslo payroll, client invoices are raised on head-office letterhead, and the 'branch accounts' consist of an annual spreadsheet allocating a share of consolidated profit to Dubai. For 2026 the spreadsheet attributes AED 2.1 million of profit to the branch; the corporate-tax return is filed on that figure.

The FTA opens a review and asks the question the spreadsheet cannot answer: how was the AED 2.1 million determined? The allocation is 8% of consolidated profit — a round number with no transfer-pricing analysis behind it, no functional analysis of what the branch actually does, and no account of the head-office services the branch consumes: engineering design performed in Oslo, the project-management software licensed centrally, the parent guarantees that win the contracts. The branch's real economics are invisible because its transactions were never recorded as the branch's transactions.

The authority's next step is predictable: it reconstructs the branch's profit itself. Without branch books, the reconstruction uses the information available — the contracts executed in the UAE, the staff physically present, the local costs that can be identified — and attributes profit on a basis far less favourable than any documented arm's-length analysis would have produced. The branch cannot rebut the reconstruction because it has no books to rebut it with; 'our spreadsheet says 8%' is not evidence. The additional tax, plus the late-payment charge running from the original due dates, exceeds the cost of five years of proper branch accounting several times over.

Now replay with separate books from day one. The branch records its own revenue (the UAE contracts), its own costs (local staff, local subcontractors, office), and the head-office charges as documented service fees under intercompany agreements priced at arm's length — the Oslo engineering support charged per hour with timesheets, the software at the per-seat licence cost, the guarantee fees benchmarked. The branch's profit of AED 1.6 million is lower than the spreadsheet's AED 2.1 million, but it is evidenced: every dirham traces to a recorded transaction, the transfer-pricing documentation supports the charges, and the functional analysis describes what the branch does and what head office does. The review, when it comes, tests the documentation and closes. The evidenced AED 1.6 million beats the unevidenced AED 2.1 million every time — because the number the authority cannot verify is not the branch's number, it is the authority's starting point.

The example's general lesson: attribution follows evidence. The branch that records its transactions controls its profit narrative; the branch that does not surrenders it. General information only — branch attribution and transfer pricing are technical areas where the facts of each case determine the outcome.

The first mistake is the notional allocation — the spreadsheet percentage with no transactional basis. Allocating consolidated profit by revenue share, headcount share, or a round number is not attribution; it is apportionment without analysis, and it collapses the moment anyone asks for the transactions behind it. Attribution starts from the branch's own recorded transactions, not from the group's consolidated result.

The second is undocumented head-office charges. The parent provides real services — management, IT, treasury, guarantees — and charges the branch amounts set by head-office finance with no agreement, no costing, and no arm's-length benchmarking. Under review, these charges face a double attack: disallowed as un-evidenced expenses of the branch, or recharacterised as something else entirely. Every recurring head-office charge needs a written intercompany agreement, a costing methodology, and a transfer-pricing analysis supporting the price — the same documentation any two independent companies would have.

The third is the invisible permanent establishment. The 'branch' is sometimes not even registered as one: a project office, a secondee arrangement, a 'liaison' presence that signs contracts and negotiates terms. The books question and the PE question are siblings — the activities that create the PE are the activities the books must record. A presence that creates taxable nexus without registration or books is the worst of both worlds: the liability exists and the evidence does not. Map the actual activities against the PE thresholds first; the books follow the footprint.

The fourth is commingled cash. Branch receipts swept daily to head office with no branch bank records, branch expenses paid from head-office cards, no branch-level cash position ever struck — the money trail that would evidence the branch's transactions does not exist. A branch bank account, used for branch receipts and payments and reconciled monthly, is the minimum infrastructure; without it, even willing accountants cannot reconstruct the branch's economics reliably.

The fifth is treating the branch as a cost centre in the group's mind. Head office sees the branch as an extension of itself — staff seconded informally, costs absorbed centrally 'for simplicity', revenue recognised wherever the contract was signed. Simplicity at head office is opacity at the branch, and opacity is what the authority taxes most aggressively. The branch must be run, in accounting terms, as if it were a separate company dealing with head office at arm's length — because for tax purposes, that is exactly what it is.

The branch's chart of accounts is the skeleton of its separate books — and it must be built as the branch's own, not as a slice of head office's. Start with the revenue lines: every UAE revenue stream separately identified — project revenue, service fees, goods sold locally — because the attribution analysis starts from what the branch earns. Then the direct costs: local staff costs (salaries, benefits, visa costs), local subcontractors, office and facility costs, local professional fees — every cost incurred for the branch's operations, recorded where incurred.

Then the head-office charges, each as its own line with its own intercompany agreement behind it: management services, IT and systems, treasury and financing, guarantee fees, royalties or licence fees where applicable. The separate lines matter because each charge faces its own arm's-length test — the bundled 'head-office allocation' line is the attribution's weakest point, concealing five different transfer-pricing analyses in one number.

The balance-sheet sections complete the picture: the branch's receivables (UAE customers), its payables (UAE suppliers), its fixed assets (the office fit-out, the equipment physically in the UAE), its cash (the branch bank account, reconciled monthly), and the intercompany balances with head office — each documented, each reconciled. The intercompany account is the most scrutinised line in any branch review: every movement evidenced, the balance agreed between branch and head-office books, differences investigated immediately.

Build the chart before the first transaction of the year, lock its structure (new accounts require approval, so the attribution analysis is not undermined by reclassification), and map it once to the transfer-pricing documentation's functional analysis — which entity does what, which charge prices which function. The chart of accounts is where the branch's separate personality lives in accounting terms; the branch that shares head office's chart, posting UAE transactions into group cost centres, has separate books in name only. Build it as if the branch were a separate company — because for the attribution, it is.

Branches are the constructed reality; the pages below cover the related risks — the bookkeeping function, reconciliations, the license lifecycle. General information only.