Assessing compliance with the arm’s length principle in transfer pricing (TP) involves conducting a benchmarking study based on high-quality comparable data. While the taxpayer can use internally available data on his transactions with unrelated parties, it’s common practice to use external data obtained from commercial databases or other sources.1 Several comparable companies are selected from a database according to certain criteria to build a range of financial results. This often raises the question of which values in that range are acceptable to demonstrate that the taxpayer’s controlled transactions are arm’s length. This article explores how wide an arm’s length range may be used in Latvia and compares how this range is interpreted in Lithuania and Estonia.
The market conditions keep changing and their analysis can never be complete because it’s not possible to find all appropriate transactions between unrelated companies. So a benchmarking study uses an approximation according to the OECD’s Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations (the “Guidelines”). Since TP is not an exact science, in many cases the benchmarking study will return a set of numbers that are all equally credible.
The Guidelines recommend that where a comparable dataset includes many observations, the credibility of analysis may be improved by using statistical tools, such as an interquartile range or percentiles, which consider the central trend and narrow the range. It’s important to note that how wide the range of values selected for comparability is may substantially affect the conclusion as to whether the taxpayer’s controlled transaction is arm’s length or whether his taxable base needs adjusting.
Latvian TP rules do not specify whether a full range or an interquartile range may be used to determine an arm’s length price, nor which value in that range is suitable for validating an arm’s length price. Yet the rules permit the taxpayer to use the Guidelines’ recommendations as an auxiliary source in a benchmarking study.
During a webinar organised by the Latvian Chamber of Commerce and Industry as part of Finance Day on 19 May 2021, the State Revenue Service (SRS) stated in its presentation:
This is also supported by the fact that we have not seen the SRS challenge a full range used in TP documentation to validate the price of controlled transactions.
It’s important to note that defining the arm’s length range is a matter of each country’s national law. Some countries may have adopted rules on a full range, while others use statistical measurements such as an interquartile range.
The table below shows conditions for using a set of comparable results to validate the price of controlled transactions in Lithuania and Estonia compared with Latvia, based on the rules and the tax authorities’ practice:
|
|
Statutory requirements |
Tax authority’s practice |
|
Latvia |
Latvian TP rules do not define an arm’s length range but state that the taxpayer may invoke the Guidelines, which allow the use of a full range. |
The SRS permits the use of a full range, including profitability indicators of all the comparable companies selected. A full range is usually taken to mean an interquartile range that allows for the minimum and the maximum value as well. |
|
Lithuania |
Lithuanian TP rules permit a pricing range that is arrived at by using statistical methods, so an interquartile range must be used. No particular quartile is defined as the most appropriate for assessing an arm’s length price. In the case of a TP adjustment, the tax authority will adjust it to the median unless it may be determined, or the taxpayer demonstrates, that a different point within the range is more appropriate. |
The tax authority adheres to the statutory definition of a quartile range. |
|
Estonia |
Estonian TP rules permit a pricing range that represents a sequence of results from the bottom quartile to the top quartile. No particular quartile is defined as the most appropriate for assessing an arm’s length price. If the TP is outside an arm’s length range, the tax authority may adjust the TP so that the result stays within the range. If the range is very wide, the analysis is repeated using more specific data or other methods. |
The tax authority adheres to the statutory definition of a quartile range. |
In benchmarking studies, we face situations where it’s difficult to find unrelated comparable companies whose functions and risks fully match the taxpayer’s controlled transactions. The main reason is that according to the NACE code the taxpayer carries out the same business as unrelated comparable companies, but in reality he performs far fewer functions and takes limited risks. In practice we increasingly see deciles, not quartiles, being used to assess comparable results in such situations.
The table below shows differences between the full range and the decile range of comparable data in assessing the same selected comparable dataset:
|
Full range |
Decile range |
|||
|
Number of observations: 42 |
||||
|
Minimum value |
1.61 |
Minimum value (0th decile) |
1.61 |
|
|
Interquartile range
|
1st quartile |
8.76 |
1st decile |
4.78 |
|
2nd decile |
7.43 |
|||
|
3rd decile |
11.48 |
|||
|
4th decile |
15.79 |
|||
|
Median |
23.69 |
Median (5th decile) |
23.69 |
|
|
3rd quartile |
33.57 |
6th decile |
28.44 |
|
|
7th decile |
32.54 |
|||
|
8th decile |
34.24 |
|||
|
9th decile |
46.19 |
|||
|
Maximum value |
88.74 |
Maximum value (10th decile) |
88.74 |
|
As we can see, using the decile range offers the taxpayer far more opportunities to assess an arm’s length price limit that best fits his limited functional profile, eliminating the difference between his own situation and that of comparable companies. The fewer functions and risks the taxpayer has, the smaller yet stronger decile he may use to arrive at an arm’s length price.
Using various statistical tools (including deciles) to improve the credibility of comparability results in assessing comparable data is endorsed by the SRS, as they often offer to use deciles when it comes to entering into an advance pricing agreement with the taxpayer.
___________________________
If you have any comments on this article please email them to lv_mindlink@pwc.com
Ask question
