Executive answer
Comparability analysis determines whether an uncontrolled transaction provides a reliable reference for testing a related-party dealing. It does not search for companies that are “similar” in the abstract. It compares the characteristics of property or services, functions, assets, risks, contractual terms, economic circumstances and business strategies under Article 179 of the Mexican Income Tax Law.
A difference does not automatically invalidate a comparable. The question is whether it materially affects price or margin and whether selection, segmentation or a reasonably accurate adjustment can eliminate its effect. If the effect cannot be measured, the comparable may need to be rejected or given less weight. Document the decision even when it supports the desired conclusion.
A defensible output includes search strategy, universe, filters, financial data, acceptance and rejection reasons, adjustments, range and limitations. Another analyst should be able to reproduce it with the same sources. A database screenshot or final list without an audit trail does not achieve that result.
Legal framework and economic role
Article 179 identifies comparability elements and requires relevant differences to be considered. Article 180 establishes methods and rules for ranges produced by comparable transactions. Comparability is therefore a condition of the method, not a decorative step after the calculation.
The OECD Guidelines provide technical direction on delineation, internal and external comparables, periods, adjustments and search processes. In Mexico they are considered according to the interpretive role recognized by the Income Tax Law where consistent with domestic law and treaties. They do not replace the statute or decide specific facts.
Intensity changes by method. A CUP requires close comparability of product and terms; a margin method may tolerate product differences but requires functional and accounting comparability. Using TNMM does not eliminate the analysis—it changes which differences matter most.
Six comparability factors
Characteristics of property, services or rights
For goods, examine quality, specification, volume, brand, stage, warranty and delivery. For services: nature, complexity, duration, personnel and output. For intangibles: rights, territory, exclusivity, life and protection. For finance: amount, currency, term, seniority, security and credit profile.
Functions, assets and risks
Compare the tested FAR, not labels such as distributor or manufacturer. An independent company owning a brand, deciding inventory and bearing warranty is not automatically comparable with an entity that merely processes sales.
Contractual terms
Review price, volume, credit, delivery, exclusivity, warranty, termination and duties. Test actual conduct. Similar written agreements may create different conditions when parties behave differently.
Economic circumstances
Geographic market, size, competition, cycle, regulation, costs, buyer power, inflation and availability affect price and margin. Do not insist on the same country without assessing alternatives, but do not combine markets without reasoning.
Business strategies
Market entry, innovation, penetration, liquidation or restructuring may explain temporary outcomes. A strategy should exist in advance and have a budget, owners and horizon. A loss does not become a strategy because the file describes it afterward.
Time and cycle
Compare relevant periods. Multiple-year data may reduce noise when the cycle supports it, but should not hide the tested year. Explain which year is tested and why contemporaneous or later-available information is used.
Internal before external comparables
An internal comparable exists when one controlled party deals comparably with an independent party or the counterparty does so. It often offers better access to terms and data, but existence does not make it reliable. Review product, volume, market, supply-chain level, credit, exclusivity and date.
A small promotional independent sale does not establish the price for a large controlled contract. A bank rate is not comparable when security, currency or term differs. A third-party service may include additional functions.
Document why an internal is accepted, adjusted or rejected. If not used, explain the material difference and why an external is more reliable. Discarding it because its answer is inconvenient undermines the analysis.
External search process
- Define the transaction and indicator before opening the database.
- Determine market, industry, functions and financial years.
- Record classification codes and search terms.
- Apply quantitative screens for economic reasons.
- Review descriptions, websites, statements and related parties.
- Exclude incompatible functions, assets or risks.
- Obtain financial data and available segmentation.
- Document acceptance, rejection and limitations.
- Test sensitivity to material screens and adjustments.
- Preserve date, database version and intermediate results.
Do not design filters to reach a target median. An objective process establishes criteria before seeing the result and explains modifications.
Acceptance, adjustment or rejection matrix
| Difference | Question | Possible action | Evidence |
|---|---|---|---|
| Product | Does it affect price or margin? | Segment, adjust or reject | Specifications and prices |
| Function | Does it add different value? | Functional selection or rejection | FAR, description and costs |
| Risk | Does it explain volatility or return? | Adjust if measurable; otherwise reject | Agreements and results |
| Market | Do demand or costs differ? | Geographic adjustment or analysis | Market data |
| Working capital | Do credit and inventory differ? | Quantitative adjustment | Balance sheets and rates |
| Accounting | Does classification alter the indicator? | Reclassify if verifiable | Notes and statements |
| Period | Is there a cycle or shock? | Justified multi-year window | Series and events |
| Intangible | Does it own unique assets? | Reject or change method | Reports and ownership |
Record materiality, reliability and effect for each decision. “Not comparable” without a specific reason is not reviewable.
If the benchmark retains only a final list, rebuild the strategy, filters, rejections and adjustments before relying on its range.
Comparability adjustments
An adjustment seeks to remove the effect of a material difference and improve reliability. It is not justified merely because a formula exists. The adjustment needs an economic relationship, consistent data and transparent method. Common examples address working capital, physical differences, volume, market or accounting.
A working-capital adjustment requires comparable account definitions, a consistent period and reasonable rate. It may be inappropriate when statements do not permit segmentation or differences reflect real functions that should not be neutralized. Document formula, direction, source, rate and sensitivity.
Do not adjust every observable difference. Each calculation adds assumptions and may reduce rather than improve reliability. Compare results before and after. If the adjustment radically transforms the range, obtain independent review.
Segmentation and accounting consistency
An independent company with several lines may be comparable only in one segment, while public data may be consolidated. Do not invent segmentation. Assess materiality, notes and business mix; reject when the tested function cannot be isolated.
For the tested party, prepare a transaction-level segmented statement with documented rules for revenue, costs, expenses, assets and liabilities. Allocation keys should relate to causation or benefit. Reconcile the segment to financial statements.
Compare accounting definitions. Gross margin, operating cost, depreciation, extraordinary items and leases may be classified differently. A consistently calculated indicator is more valuable than a larger sample with incompatible metrics.
Comparability in changing markets
Inflation, tariffs, disruption, capacity and volatility may affect comparables and the tested party differently. Do not remove a year automatically or use averages to erase an event. Identify the cause, the party controlling the risk and observable data.
Consider information lag. Contemporary financial statements may not be available when the report is prepared. Document what was available, whether prior years were used and how the set will be refreshed. A refresh policy may combine periodic full searches with annual independence and results checks while facts and criteria remain stable.
Common errors
- Selecting companies by industry without FAR.
- Ignoring an internal comparable.
- Changing filters after seeing the range.
- Relying only on commercial descriptions for qualitative review.
- Mixing consolidated and segmented statements.
- Applying adjustments without an economic relationship.
- Using multiple-year averages to hide the tested year.
- Failing to document rejections.
- Keeping screenshots instead of reproducible outputs.
- Treating the range as proof of materiality or deductibility.
Quality control
An independent review should reproduce the population and calculations, test duplicates, verify related-party exclusions, inspect filters and recalculate indicators and adjustments. Compare the result with FAR, agreements and performance. Investigate outliers and changes from the prior year.
Record limitations: unavailable segmentation, broad geography, few comparables, lagged data or imperfect adjustments. Transparency does not weaken the report; it clarifies the proper weight of the evidence.
Benchmark audit trail and governance
Assign every candidate an identifier and retain source, date, screen, reviewer, decision and reason. When a decision changes, preserve the earlier version and new evidence. This trail explains why a company entered or left the range without depending on an analyst’s memory.
Define who may change criteria, approve adjustments and close the sample. Material changes after results are calculated need a second review. Compare the set with the prior year as well: additions, removals, mergers, losses, functional changes and financial availability.
A dashboard may show the initial universe, candidates by screen, accepted companies, rejection categories and the effect of each adjustment. It should not turn judgment into an automatic score. Its role is to make the process visible and flag inconsistent decisions.
Before using the range for a true-up, confirm that the tested-party result uses the same period, currency, classification and indicator definition. A sound benchmark applied to the wrong segment produces a misleading conclusion with quantitative precision.
Related topics
- PT-017: FAR analysis.
- PT-019: method selection.
- PT-020: range and median.
- PT-021: internal and external CUP.
- PT-025: TNMM, tested party and indicator.
Acceptance-and-rejection log
For every comparable, record source, search date, activity, market, period, independence, financial availability, differences and decision. A rejection should identify the economically relevant fact rather than use generic wording. Acceptance also needs a reason. Retain original search results so another person can reproduce the population before filters.
Assess functions, assets, risks, products, geography, market level, agreements, volume and business cycle. Decide whether a difference prevents comparison, can be controlled through selection or permits a reliable adjustment. Document the formula, data and sensitivity for every adjustment; do not improve statistical appearance without economic support.
A second reviewer challenges outliers, filters and year-to-year consistency. Where the market changed, update the search instead of retaining companies for convenience. The log connects economic judgment to evidence and explains why the final set is more reliable than seemingly similar alternatives.
Sources and verification date
- Mexican Income Tax Law, current text, Articles 179 and 180.
- OECD Mexico Transfer Pricing Country Profile.
- OECD Transfer Pricing Guidelines, a technical reference.
Sources checked on August 2, 2026. Reliability depends on facts, method, data and period; a commercial database does not replace technical judgment.
Request a Comparability Review to assess internal comparables, searches, screens, rejections, segmentation, adjustments and benchmark reproducibility.