Quick Definition
Interquartile Range (IQR) — Interquartile Range (IQR) is a statistical measure used in transfer pricing to define the arm's length range by excluding the top and bottom 25% of comparable results.
Our glossary is continuously expanding. For deep-dive guides on specific topics, explore our blog or contact us with suggestions.
Interquartile Range (IQR) is a statistical measure used in transfer pricing to define the arm's length range by excluding the top and bottom 25% of comparable results. The IQR spans from the 25th percentile (Q1) to the 75th percentile (Q3), capturing the middle 50% of data points. It is the default range measure preferred by most tax authorities because it mitigates the impact of outliers and imperfect comparability.
Formula:
Where:
The OECD Transfer Pricing Guidelines (2022) endorse the interquartile range in Chapter III at as part of the guidance on comparability analysis. The Guidelines note that where comparability defects remain in a comparable set, it may be appropriate to use measures of central tendency (such as the median) to minimize the risk of error from unknown comparability defects.
US Treasury Regulations §1.482-1(e)(2)(iii)(B) provide guidance on using the interquartile range, stating that the arm's length range is derived from the results of all uncontrolled comparables that achieve a similar level of comparability.
The IRS has consistently used the interquartile range in practice and in APAs to address comparability defects.
The IQR is the default approach for establishing arm's length ranges in most transfer pricing analyses. It serves two purposes:
When the IQR is appropriate:
When the full range may be acceptable:
Jurisdictional Variation: While most tax authorities prefer IQR, calculation methods differ. The US/IRS uses a specific interpolation method, while OECD guidance is less prescriptive. Always verify the applicable method for your jurisdiction.
Comparable Set: 9 companies with the following Operating Margins:
| Company | Operating Margin |
|---|---|
| A | 1.2% |
| B | 2.1% |
| C | 2.8% |
| D | 3.5% |
| E | 4.0% |
| F | 4.6% |
| G | 5.3% |
| H | 6.8% |
| I | 9.5% |
IQR Calculation (OECD Inclusive Method):
With 9 data points:
Arm's Length Range:
Interpretation: A tested party with Operating Margin between 2.45% and 6.05% falls within the interquartile range. Results outside this range may require adjustment or additional analysis.
Different jurisdictions use different interpolation methods:
| Jurisdiction | Method | Q1 Formula |
|---|---|---|
| OECD (general) | Inclusive | 0.25 × (n + 1) position |
| US/IRS | Interpolation | Specific interpolation per Reg. §1.482 |
| India | Arithmetic | Simple 25th percentile |
| UK | Flexible | Pragmatic approach based on dataset |
| Germany | Median focus | Often emphasizes median over range |
Practical Tip: For a detailed walkthrough of IQR calculation methods by jurisdiction with worked examples, see our IQR Calculation Methods guide.
| Factor | Use IQR | Use Full Range |
|---|---|---|
| Comparability quality | Imperfect but reasonable | Near-perfect |
| Sample size | Larger sets (7+) | Smaller, carefully vetted sets |
| Outlier presence | Likely or unknown | No outliers, all justified |
| Tax authority preference | Most jurisdictions | Specific jurisdictions allow |
| Audit defensibility | Safer default | Requires strong justification |
The IQR removes the top and bottom 25% of results, which mitigates the impact of outliers and imperfect comparability. Even with careful screening, comparables aren't perfectly identical to the tested party. Outliers may result from unobserved factors (one-time events, different accounting treatments, hidden risks). The IQR provides a more reliable, defensible range by focusing on the central tendency.
If the tested party's result falls outside the IQR, most tax authorities can make an adjustment to bring the result to a point within the range—typically the median. suggest the median minimizes error risk. Falling outside the IQR doesn't automatically mean non-compliance, but expect scrutiny and potential adjustments during audits.
While there's no universal minimum, 7-10 comparables is generally considered sufficient for a meaningful IQR. With fewer than 5 data points, statistical measures become less reliable. Some practitioners accept 5-7 with strong justification. More important than quantity is quality—8 well-screened comparables beats 20 loosely comparable ones.
Yes, significantly. Different interpolation methods (OECD inclusive, US interpolation, simple arithmetic) can produce different Q1 and Q3 values from the same dataset. The difference is typically small but can matter when the tested party is near the boundary. Always use the method accepted in your jurisdiction and document it clearly.
For multi-year benchmarking, you can either: (1) calculate IQR for each year separately, or (2) calculate weighted-average results per comparable first, then compute IQR from those averages. The latter is more common and reduces year-to-year volatility. Ensure your approach is consistent and documented.
Only with legitimate, documented reasons (failed manual screening, discovered facts). Never exclude comparables solely because they widen your range unfavorably. Tax authorities view selective exclusion skeptically. If a comparable passes your screening criteria, it should remain in the set—the IQR mechanism handles outliers automatically.
The IQR is not a statistical confidence interval in the traditional sense. It doesn't represent a 50% probability that the "true" arm's length price falls within the range. Rather, it's a practical convention for defining acceptable results when perfect comparability is impossible. The IQR captures the "central tendency" of comparable outcomes, acknowledging that extremes may reflect comparability defects.