10 Documentation Weaknesses That Trigger Audits (and How to Fix Them)

The weaknesses below are the recurring findings in transfer pricing examinations — not the exotic failures, but the ones that appear file after file, year after year, in the TPO’s orders and the AO’s assessments. Each is a gap between what the documentation standard requires and what the file actually contains, and each has a remediation that is a practice, not a project. Use the table as the pre-filing documentation audit: each row is a question the file should be able to answer with evidence.

#	Weakness	The risk in examination	The remediation
1	Documentation prepared after the notice — the file assembled in response to the 282BC notice, not within the 31 May window	The penalty shield is down: section 271AA’s 10% penalty is live regardless of the pricing outcome, and the “contemporaneous” character is indefensible on the dates	The study runs on the year’s actuals before 31 May; the preparation dates are evidenced (the working file’s chronology, the sign-off). See contemporaneous documentation
2	The method not stated — or changed silently — the Local File’s method claim missing, or the method/PLI different from the prior year without explanation	The method-choice question becomes an open question the TPO answers: the method substitution lands on a file that never documented its choice	The method register in the TP policy; the Local File’s method block states the method, the PLI and the choice rationale; any year-on-year change carries the functional-change reason
3	The PLI definitions inconsistent — the benchmark working’s operating cost differs from the Local File’s, which differs from the return’s	The TPO re-derives the PLI on its own lines and the position moves; the inconsistency is citable line by line	One cost definition, used in every document, stated in the file; the reconciliation of the PLI across the benchmark working, the Local File and the return is a named check
4	The matrix without reasons — accepts and rejects with no specific, evidence-grounded rationale per company	The pool is a list, not a method: the TPO rejects the matrix on credibility and rebuilds the pool on its own filters — the range the file defended stops existing	Every disposition carries its grounded reason; overrides carry their own rationale; see defending the matrix and the 12 benchmarking mistakes
5	The cost base undefined — the mark-up benchmarked on a cost base that is never shown, line by line	The cost-base question is the cost-plus file’s entire defence, and an undefined base has no defence: the TPO’s base is the base	The line-by-line cost base in the file, with each line’s definition, consistent across years and documents — the contract manufacturing discipline applied to every cost-based method
6	The segment data missing — consolidated financials used for a segment function, with no segment split	The comparable is not comparable (the function is buried in the group’s mix), and the tested party’s own segment is not isolated: the pool and the tested party are both misstated	The segment (or the related-party split) extracted and documented for the function; where the segment data is genuinely unavailable, the exclusion of the false comparable is reasoned, not silent
7	The agreement contradicts the conduct — the intercompany paper describes functions, IP or pricing the conduct does not show (or vice versa)	The characterization follows the substance: the file that presents the agreement as the whole story is presented an incomplete story, and the examination reads the conduct	The consistency test before filing (agreement vs conduct, per examined fact); the amendment trail where the conduct evolved; see intercompany agreements
8	The three tiers do not reconcile — the Master File, the Local File and the CbCR carry different numbers for the same facts	The inconsistency is the finding that needs no adjustment to be damaging: it is the risk assessment’s input, the CbCR flag’s confirmation, and the credibility collapse across the documentation set	The named reconciliation step in the documentation cycle: the structure, the segment data, the related-party revenue, the tax positions — reconciled across the three tiers and the statutory accounts, documented
9	The return deviates from the documentation — the TP position in the return differs from the documentation’s position	Section 271AA’s second limb: the penalty is live on the inconsistency between the return and the file, even where the file itself is contemporaneous	The return-in-accordance check as a named step: the prices, the characterizations and the computations in the return verified against the documentation before filing
10	The documentation not maintained — the working file (the matrix evidence, the overrides, the segment workings) unavailable in year N+3, because the retention practice kept the Local File and lost the substance	The production standard does not relax with age: the 8-year retention covers the documentation and the working file behind it, and the file that cannot be produced in year five is the file that does not exist for the examination	The retention discipline on the full working set (the search design, the matrix with evidence, the overrides, the segment workings, the reconciliations) — maintained, indexed and retrievable for the full 8 years

The audit, in order

Work the list top to bottom against the year’s file:

1. The dates — prepared within the window, the dates evidenced, the production ready (row 1).
2. The method — stated, reasoned, consistent year on year (row 2).
3. The definitions — one cost definition, one PLI, across every document (row 3).
4. The matrix — every disposition reasoned, every override rationaled (row 4).
5. The cost base — line by line, defined, consistent (row 5).
6. The segments — the function isolated, the false comparables excluded with reasons (row 6).
7. The paper — the agreements matching the conduct, the amendments current (row 7).
8. The tiers — the three documentation tiers reconciled to each other and to the accounts (row 8).
9. The return — filed in accordance with the documentation (row 9).
10. The retention — the full working set maintained for the eight years (row 10).

A file that passes all ten is a file that presents as the application of a standard — and the examination of a standard’s application is a different, and winnable, fight from the examination of a year’s reconstruction. A file that fails any of the ten knows, before the notice, which finding will arrive first.

See also

• TP audit defense
• Transfer pricing policy
• Defending the Accept-Reject matrix
• The documentation pillar guide