Art and Science of Benchmarking Analysis: Strategies to Avoid "Cherry Picking" Traps in Transfer Pricing

In the international tax ecosystem, Benchmarking Analysis is the backbone of proving the fairness of an affiliated transaction. This analysis is not merely a statistical data search, but a systematic process to find comparable independent transactions to determine whether the price or profit in the affiliated transaction has met the Principle of Fairness and Business Prevalence (Arm’s Length Principle).

However, the biggest challenge in this process is subjectivity leading to cherry picking—the practice of selectively choosing comparables that only favor the Taxpayer's position and ignoring other data that might be more relevant but less favorable.

The Essence of Benchmarking Analysis

Benchmarking is the process of comparing the conditions of an affiliated transaction (controlled transaction) with the conditions of a transaction carried out by an independent party (uncontrolled transaction). The goal is to find a meeting point called "comparable".

An independent transaction is considered comparable if there are no material differences affecting price or profit, or if such differences exist, reasonably accurate adjustments (reasonably accurate adjustment) can be made to eliminate the impact of those differences.

Data Search Approaches

In practice, there are two main approaches in searching for comparables:

1. Additive Approach: The analysis preparer compiles a list of third parties believed to carry out comparable transactions, then collects information to confirm it. This is often used if the analyst already possesses in-depth market knowledge.
2. Deductive Approach: Starts with a broad set of data (for example from a commercial database), then filters it using quantitative and qualitative criteria to obtain a final set of reliable comparables.

The "Cherry Picking" Phenomenon: Enemy of Data Integrity

Cherry picking occurs when the Taxpayer or tax authority selectively chooses data that supports their position and discards data that does not support it, without objective reasons.

A classic example is rejecting comparable companies having low profit levels on the grounds of "inefficiency" without further analysis, merely so that the arm's length range (arm's length range) becomes higher, or vice versa. This practice damages the credibility of transfer pricing documentation (TP Doc) and often becomes a primary target for correction by tax authorities.

Golden Rules to Avoid Cherry Picking

To ensure benchmarking results can be defended before tax authorities (such as DGT in Indonesia or IRBM in Malaysia), Taxpayers must adhere to the following rules:

A. Transparency and Reproducibility (Reproducibility)

The comparables search process must be transparent, systematic, and verifiable. If another party conducts a search using the same criteria and database, they should obtain the same results. Inclusion or exclusion criteria for data must be documented in detail in the TP Doc.

B. Hierarchy of Comparable Selection

Regulations give priority to Internal Comparables (transactions between the Taxpayer and an independent party) over External Comparables. Ignoring available internal comparables in favor of using more favorable external comparables is a fatal form of cherry picking.

C. Quantitative Screening Strategy (Quantitative Screening)

The use of diagnostic ratios to filter comparables is permitted, but must be objective. Inclusion criteria such as data availability, independence (related party transactions < 25%), and industry code suitability (SIC/NACE) must be applied consistently.

D. Qualitative Review (Qualitative Screening/Manual Review)

This is the final line of defense. The candidate list must be examined manually (annual reports, websites) to ensure comparability of functions, assets, and risks. Rejection must be based on material reasons such as differences in functional profiles or ownership of unique intangibles.

E. Use of Arm’s Length Range (Arm’s Length Range)

Because transfer pricing is not an exact science, results are usually in the form of a range of numbers. The use of the Interquartile Range (Q1 to Q3) is highly recommended—and mandatory in Indonesia if there are more than two comparables—to discard outlier data (outlier).

F. Consistency (Consistency)

Search strategies must be consistent from year to year unless there are material business fact changes. Changing search criteria only to obtain a "good" set of comparables is unacceptable.

G. Diagnostics and Cross-Check

Use diagnostic ratios (for example the ratio of marketing costs to sales) to filter function intensity levels to achieve an "apple to apple" comparison.

Benchmarking analysis is an exercise in objectivity. The key to avoiding cherry picking allegations is strong documentation explaining "why" and "how" a piece of data was selected or rejected. Analysis showing a transparent, logical, and consistent selection process—even if the results are less favorable—is far more valuable and defensive than manipulated analysis. This is the Taxpayer's greatest asset in the era of global tax transparency.

References

1. OECD. (2022). OECD Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations 2022. OECD Publishing, Paris.
2. United Nations. (2021). Practical Manual on Transfer Pricing for Developing Countries (2021).
3. Ministry of Finance of the Republic of Indonesia. (2023). Regulation of the Minister of Finance of the Republic of Indonesia Number 172 of 2023 concerning the Application of the Principle of Fairness and Business Prevalence in Transactions Influenced by a Special Relationship.
4. Directorate General of Taxes. (2013). Regulation of the Director General of Taxes Number PER-22/PJ/2013 concerning Guidelines for Audit of Taxpayers Having a Special Relationship.
5. Inland Revenue Board of Malaysia. (2024). Malaysia Transfer Pricing Guidelines 2024.