Benchmarking Studies

The benchmarking study is the specific analytical exercise that establishes an arm's length price or margin range for a related-party transaction, using external comparable company or transaction data. The Local File is the broader document — business description, transaction detail, FAR analysis, method selection — that the benchmarking study sits inside as supporting evidence. A Local File can theoretically exist without a benchmarking study only where the method chosen (an internal CUP, for example) does not require external comparables; most Local Files rely on at least one benchmarking study for their TNMM- or Cost Plus-based transaction categories.

Only where the pricing method selected for that transaction category actually requires external comparable data. Where a genuinely reliable internal comparable exists — the same product or service sold to an independent party on comparable terms — CUP using that internal comparable can be more direct and does not need a full external database search. Low value-adding intra-group services may also qualify for a simplified fixed-markup approach under OECD guidance rather than full benchmarking.

A commercial transfer pricing database aggregates standardised financial data on a large population of independent companies, searchable by industry classification, geography, size, and independence status — the scale and standardisation needed for a defensible search. Free public sources are typically incomplete, inconsistently formatted, and lack the independence and related-party revenue flags that quantitative screening depends on, making a credible, reproducible search difficult to run from them alone.

The tested party is generally the party to the transaction with the simpler, less complex functional profile — fewer unique intangibles, less risk, more routine functions — because its results are easier to benchmark reliably against a population of independent comparables performing similar routine functions. It is not automatically the UAE entity; where the foreign counterparty is the simpler party, benchmarking may need to reference that entity's results instead.

Standard quantitative filters remove companies that fail objective, database-searchable criteria: those with a controlling related-party shareholder (failing independence), companies below a minimum revenue or asset size, companies with persistent losses over the review period (often considered unreliable comparables), and companies whose own related-party revenue exceeds a set proportion of total revenue, since their reported margins may themselves reflect non-arm's length pricing.

Qualitative screening is the manual review of each company's actual business description against the tested party's real functional profile — checking that a company which survives the quantitative filters on paper genuinely performs comparable functions, not just that it sits in a similar industry code. Industry classification codes are often too broad to capture true functional comparability, so this step catches companies that look right numerically but are functionally different.

The interquartile range narrows the full set of comparable results to the middle 50%, removing the highest and lowest quartiles — the outliers most likely to reflect an imperfect comparable match rather than genuine market variation. Using the interquartile range is a widely accepted way to improve the reliability of the arm's length range where perfect comparability cannot be achieved, which is the normal case in practice.

A result outside the range does not automatically mean an adjustment is required, but it does mean the position needs closer examination — reviewing whether the comparable set or the tested party's own data needs revisiting, whether an adjustment to the pricing is warranted, or whether there is a documented, defensible reason for the outcome (a start-up loss period, a one-off cost, a genuine market condition). PNPC flags this to the client immediately rather than silently reporting a range that does not support the current pricing.

Financial data underlying the study is generally refreshed annually to reflect the tested party's actual results for the current year. The underlying comparable company search itself is typically refreshed on a multi-year cycle, commonly every three years or sooner if the industry or the tested party's functional profile changes materially, consistent with OECD guidance on documentation currency.

Sometimes as a starting reference, rarely as a direct substitute. A UAE-specific benchmarking analysis needs to reflect the UAE tested party's own functional profile and, ideally, comparable companies operating in a broadly similar market, which a global template built around a different entity's functions and geography rarely captures accurately. We review any existing group-level study first, but typically still need to run or adapt the search specifically for the UAE entity.

No, generally different sources are used. Owner and director remuneration is benchmarked against salary and compensation survey data appropriate to the role, industry, and market; related-party premises rent against comparable market rental listings; shareholder loan interest against market lending-rate references. These are market-reasonableness checks under Article 36 rather than a transfer pricing database search of independent company financials.

Where the population of genuinely comparable independent companies is small — a narrow or specialised industry, a jurisdiction with limited public financial disclosure — the search is typically broadened geographically or by relaxing selected quantitative criteria in a documented, defensible way, rather than accepting poor-quality comparables to hit a target sample size. The reduced sample size and the reasoning for broadening the search are both recorded transparently in the study.

Where a Free Zone entity's related-party transactions are covered by this study, the benchmarking outcome directly informs whether the pricing supports the arm's length condition underpinning Qualifying Income classification. A related-party transaction priced outside the defensible arm's length range risks recharacterisation, which can in turn affect whether the associated income still qualifies for the 0% rate, or in more serious cases whether the non-qualifying revenue threshold for QFZP status is breached.

For a single transaction category with the tested party and method already confirmed, a realistic timeline is three to five weeks from search kickoff to a finalised, internally reviewed range — covering database search, quantitative and qualitative screening, financial data extraction, and range computation. Multiple transaction categories in the same engagement, or a first-time build that also requires FAR analysis and method selection, extend the timeline.

Yes. For groups with related-party flows between an Indian and a UAE entity, PNPC's India and Dubai desks work from a single agreed set of facts about the transaction and each party's functional profile, so the tested party characterisation and, where feasible, the comparable logic are consistent across the UAE benchmarking study and the Indian transfer pricing documentation supporting Form 3CEB.

If an FTA reviewer challenges specific comparables or the overall range during an information request or audit, PNPC engages directly on the technical merits — explaining the search methodology, the screening logic applied, and why the accepted comparables genuinely reflect the tested party's functional profile. Where the FTA proposes an adjustment we consider unsupported, we assist with a formal reconsideration request or escalation through the applicable dispute resolution route.

PNPC can scope benchmarking as a standalone engagement — for example, a refresh of an existing comparable set, or a study for a single new transaction category — or bundled into a full first-time Local File build where multiple transaction categories each need their own benchmarking support. Pricing depends on the number of transaction categories, the availability of the tested party's financial data, and whether existing group-level studies can be adapted rather than built from scratch.

PNPC searches recognised commercial transfer pricing databases on the geographic scope that is genuinely defensible for the tested party's industry and market — in practice this is rarely confined to UAE-incorporated companies alone, since the population of financially disclosing independent UAE companies in most industries is too small to build a statistically reliable range from. The search is more commonly run across the GCC, wider MENA, or a broader Asia-Pacific/global set depending on what the tested party's own market genuinely looks like, with the geographic broadening documented and reasoned rather than left unexplained.

The screening discipline is the same, but the functional match needs particular care. A Free Zone entity benefiting from Qualifying Free Zone Person 0% treatment is often a leaner, more narrowly-functioned business than a fully taxed mainland operating company in the same broad industry classification, so comparables drawn from larger, full-risk mainland-style operators can overstate the margin the Free Zone entity's actual functions would earn at arm's length. We test functional fit particularly closely for Free Zone tested parties for this reason.

Generally several, run as parallel workstreams within a single engagement. Each transaction category — say, an intercompany management fee and a separate royalty arrangement — typically has its own tested party logic, comparable population, and method, since a distribution margin benchmark and a royalty rate benchmark draw on entirely different comparable universes. We scope and price each category's benchmarking component distinctly even where they are delivered together.

An intercompany loan is typically benchmarked using CUP against a market interest-rate reference appropriate to the borrower's credit profile, tenure, currency and any security — not against a database of independent companies' profit margins. An intercompany guarantee fee, where a parent guarantees a UAE subsidiary's third-party borrowing, is benchmarked against the value of the credit enhancement the guarantee actually provides, typically referenced to the difference between the borrower's stand-alone and guaranteed borrowing terms where that data is available. Both use a market-reference approach rather than the commercial company database used for TNMM or Cost Plus.

Not automatically. A single loss year can be consistent with genuine market conditions — a start-up period, a one-off cost, a sector-wide downturn — and does not by itself invalidate the benchmarking approach. What it does trigger is closer scrutiny: checking whether the loss reflects the tested party's own operational reality rather than a mispriced related-party transaction, and whether the comparable set itself includes persistent loss-makers that should have been screened out under the quantitative filters.

Sometimes, where the two entities genuinely share the same functional profile, industry, and transaction type — in which case a single comparable search can support both tested parties' ranges, with each entity's own financial result separately tested against the shared range. Where the entities differ in scale, risk profile, or market even slightly, we run separate searches rather than force a shared range onto functionally distinct entities.

Practice varies by method and industry, but using multiple years of comparable company data — commonly three years — is a widely accepted approach under OECD guidance to smooth out year-to-year volatility that might otherwise make a single comparable year an unreliable reference point, particularly for TNMM-based ranges. We select the approach appropriate to the method and industry rather than defaulting mechanically to either a single-year or multi-year basis.

We generally hold the study's tested-party result comparison until the audited or at least finalised management figures are available, since benchmarking against provisional numbers that later move is inefficient work that usually needs revisiting. Where a filing deadline genuinely cannot wait, we can finalise the comparable search and range on schedule and slot in the tested party's confirmed result once available, but we flag this as an interim approach rather than treat it as the finished study.

Resale Price Method is generally more reliable where gross margin data for genuinely comparable independent distributors performing a similar function is available and accounting policies (how cost of goods sold is classified) are reasonably consistent across the comparable set. Where that gross-margin-level comparability is hard to establish — differing accounting treatments across jurisdictions, or limited disclosure at the gross margin level — TNMM's net margin basis is often more robust in practice because it is less sensitive to those classification differences. The choice follows from the FAR analysis and the data actually available, not a fixed rule.

The benchmarking mechanics — search, screening, range computation — are the same, but the tested party characterisation needs particular care for a branch, since a branch is legally part of the same entity as its foreign head office rather than a separate related party. Transactions between a head office and its own branch are typically analysed under profit attribution principles rather than as a related-party transaction requiring external benchmarking in the same way as dealings between two separately incorporated related entities — though genuine related-party dealings between the branch and other group entities are still benchmarked in the usual way.

Not automatically — the benchmarking methodology itself (search, screening, range computation) is generally unaffected by a tax rate or threshold change, since it tests whether pricing is arm's length, not what rate applies to the resulting profit. What can change is whether the transaction category still requires documentation at all, or whether a Free Zone entity's Qualifying Income position is affected — both of which we reassess whenever Ministry of Finance or FTA guidance is updated.

Yes, and this is generally the stronger approach where possible — benchmarking a proposed royalty rate, management fee structure, or intercompany loan before it goes live means the pricing is set with the benefit of a proper comparable search from the outset, rather than tested retrospectively against a price already chosen for other reasons. The search methodology is the same; the difference is timing and that the study supports a forward pricing decision rather than defending a historical one.

No — the absence of industry-specific FTA guidance does not reduce the Article 34 arm's length obligation, it just means the benchmarking study leans more heavily on general OECD Transfer Pricing Guidelines methodology and sound professional judgement in applying it, since there is no sector-specific shortcut to fall back on. If anything, a well-documented, transparent screening trail matters more, not less, where there is less published guidance for a reviewer to reference alongside the study.

There is no fixed minimum — the right question is proportionality, not a hard cut-off. For a transaction genuinely below the disclosure or documentation threshold, a full external comparable search is unlikely to be the proportionate response, and a lighter qualitative rationale may suffice; PNPC scopes this on a case-by-case basis rather than applying a blanket size rule, since aggregated small transactions in the same category can still cross a threshold collectively even where none does individually.

The two can run in parallel — benchmarking depends on the related-party mapping and FAR analysis, not on Corporate Tax registration being finalised first — but the Local File and disclosure form the benchmarking feeds into do reference the taxable person's Tax Registration Number, so we make sure registration is confirmed before the documentation is finalised for filing, even if the underlying search work started earlier.

Almost entirely remotely — the engagement is a documentation and analysis exercise built on financial data, agreements, and functional interviews conducted by video call, database research, and document exchange, without a physical-presence or biometric step. The one genuine dependency is timely access to someone on your finance or operations team who can speak accurately to what the tested party actually does, since the FAR conversation that underpins the whole search cannot be reconstructed from documents alone.

The full range is every accepted comparable's result, unadjusted. The interquartile range narrows that to the middle 50%, and the median is the single midpoint value within it. PNPC reports the full range, the interquartile range, and the median together, with the tested party's actual result placed against all three — because a reviewing officer may look at any of the three depending on the specific point being tested, and presenting only one narrows the study's usefulness unnecessarily.

The historical study remains a valid record of the pricing analysis performed for the periods it covers, but a change in ownership or group structure is a trigger to reassess the tested party, the related-party map, and potentially the comparable set going forward — since the counterparties, the functional profile, or even which transactions are related-party at all can change materially through an acquisition or restructuring.

Where practical and where reliable data supports it, yes — running a secondary method as a sense-check on the primary method's result (for example, a rough Cost Plus cross-check alongside a primary TNMM analysis) can strengthen the study's credibility, though it is not always feasible or necessary depending on data availability and the transaction's complexity. We apply this selectively rather than as standard practice on every engagement, since a secondary method run on weaker data can undermine rather than reinforce the primary result.

We tell the client directly and explain the reasoning, rather than benchmarking to the historically used method by default. A group may have selected a method years ago for reasons unrelated to a proper FAR analysis — administrative convenience, a template from another jurisdiction, or a method a previous advisor simply defaulted to — and if the current FAR analysis points to a better-fitting method, continuing with the old one purely for consistency produces a weaker study than switching, with the change and its rationale documented.

A well-documented benchmarking study — search methodology, screening criteria, rejection matrix, and range computation — is the primary evidence used in a Mutual Agreement Procedure (MAP) discussion between the two tax authorities where the UAE has a Double Taxation Avoidance Agreement with the counterparty jurisdiction, since MAP negotiations turn substantially on whose economic analysis is more defensible. A thin or unscreened study leaves the UAE side of that negotiation weaker from the outset.

PNPC scopes benchmarking engagements as a fixed, written fee agreed after an initial scoping call, based on the number of transaction categories, whether it is a first-time build or an annual refresh, and the complexity of the comparable search required — not billed as open-ended hourly time against a deliverable with a defined analytical scope.

Yes, though the tested party side of the analysis relies on whatever partial-year or forecast financial data is available rather than a full prior-year track record, and the study should be explicit about that limitation. The comparable search and screening methodology work the same regardless of the tested party's own operating history, but we flag a first-year study's tested-party comparison as provisional where the entity's own results are genuinely incomplete or unrepresentative of a full operating year.