The Firm

Show Timings:

Friday: 10.30 pm, Saturday: 11.30 am

Sunday: 9:30am & 11.00pm


Deciphering A 'Range' Of Transfer Pricing Issues

Published on Thu, May 28,2015 | 21:34, Updated at Thu, May 28 at 21:34Source : 

By: Keyur Shah, Partner – Tax & Regulatory services, EY

Since the time the transfer pricing (TP) provisions were first introduced in India in 2001 and the concept of the arm’s length price introduced, taxpayers have been grappling with issues arising from the requirement to use the arithmetic mean to defend the arm’s length nature of their transactions.

The TP provisions as they currently stand require that, where more than one arm’s length price is arrived at, the arithmetic mean of all such prices shall be the arm’s length price. Limited flexibility is extended to taxpayers by allowing a tolerance band of 3 percent of the transaction price (1 percent in the case of wholesale traders).  

Tax experts have advocated the introduction of the ‘range’ for computing the arm’s length price, in line with the developed economies across the globe, since a range of arm’s length prices would be more in line with commercial realities faced by a taxpayer. The OECD Guidelines also recognise the range concept.  

Representations made received a fillip when the Finance Minister announced as a part of his first budget, the intention to permit the use of the range and multiple year data for determining the arm’s length price, based on rules to be prescribed. On 21 May 2015, the CBDT rolled out the draft rules relating to the use of the range and multiple year data for transactions undertaken on or after 1 April 2014, and has invited comments from stakeholders by 31 May 2015.

The use of the range and multiple year data have been made applicable only in case of gross and net profit margin based methods (i.e. Cost Plus, Resale Price and Transactional Net Margin Methods), and it is difficult to understand the rationale behind the CBDT’s proposal to restrict the applicability to only these methods.

The Range
The proposed conditions for the use of the ‘range’ are:

·         At least nine comparable companies to be selected, having data for at least 2 out of the 3 year period consisting of the relevant financial year plus previous 2 years, and weighted average to be computed

·         Data points lying between the 40th and 60th percentile would constitute the range

·         In case the transaction price falls outside the range, the median of the range to be considered to compute a TP adjustment

Practically, the requirement to select at least nine comparable companies for adopting the ‘range’ seems to ignore the ground realities of many industries where availability of quality data in the public domain on comparable companies is a challenge. For such industries, the range concept as it is currently envisaged would be impossible to opt for. Since a higher number of comparables would not always ensure a more reliable range the rationale behind using atleast nine comparables is difficult to fathom. Most developed nations which permit the use of a range to benchmark their transactions do not contain such a threshold.

The draft rules provide for the range to be computed as the 40th and 60th percentile of the data points. The 40-60 percentiles are a departure from the internationally recognised concept of the inter-quartile range i.e. the 25-75 percentiles, mandated in countries such as USA and Singapore, and commonly followed in practice in other countries such as UK and Australia. The 40-60 percentiles could actually prove restrictive vis-à-vis the arithmetic mean with the 3% tolerance band, where the comparable margins do not vary significantly.

However, the manner of computation of the 40-06 percentile has not been provided for in the rules. For instance, the manner of computation using MS Excel would provide a different result as compared to traditional statistical techniques.

Further, the rules provide that if the transaction price of a taxpayer falls outside the range, the median of the range would be considered for computing the TP adjustment. It is not clear whether the median would be the 50th percentile or an average of the 40th and 60th percentile or whether a median of the data points falling with the 40th and 60th percentile. For instance, as per the guidance provided by the US IRS, the median of all the data points is considered for the purposes of making an adjustment.

In view of the different approaches to computing the percentiles and the median, clarity needs to be provided on the computational methodology in the final rules to avoid future litigation.

The current provisions of the use of the arithmetic mean shall continue to apply in cases where the arm’s length price is being determined by using the CUP, PSM or the Other Method, or where there are less than nine comparable companies being identified.

Use Of Multiple Year Data
The current TP provisions specify that the data to be used in analysing a comparable company should be the data relating to the financial year in which the transaction has taken place. Further, where relevant, the data relating to the two preceding years can also be used.

Various court decisions have interpreted the term ‘data’ used in the above provisions to refer only to the financial information/ profit margins of the comparable uncontrolled enterprises, and have accordingly held that a taxpayer should use only the financial data of the comparable companies relating to the financial year in which the transaction has taken place, unless the taxpayer is able to clearly demonstrate that the use of the data relating to the preceding two years is relevant.

Accordingly, the existing provisions are not meant to allow the use of multiple year data as a norm i.e. an average of the margins of current and previous two years, but use the data of previous years only to establish comparability.

The proposal to use multiple year data is a welcome move as obtaining adequate current year data of comparables at the time the tax payer filed its return of income was practically impossible. Further, use of multiple year data could capture market cycles and reduce distortions in financial results arising from the use of single year data. The conditions in the draft rules are -

a)      The multiple year data shall include three years i.e. the year in which transaction has been undertaken (current year) and previous two years.

b)     However, the use of any two years data out of three years is permitted in case the data for the current year is not available in the databases at the time of filing the return of income, or if a comparable fails to qualify for a quantitative filter (eg; related party or export filter) in a particular year, or if a comparable has commenced operations in last two years or closed down operations in the current year.

c)      The use of current year data, which was not available at the time of filing the return of income, has been permitted to be used by both the income-tax department and / or the assessee if the same becomes available subsequently, at the time of audit.

The CBDT could consider allowing data for more than two years to be used. The Australian Tax Office permits the use of multiple year data for five years, i.e. the current year data and preceding four years.

Further, TP regulations in the US also do not restrict the use of multiple-year data to a specific time period.

The use of multiple-year data has been proposed to be made mandatory in case the arm’s length price is determined by using the RPM, CPM or TNMM.  

In the use of the multiple year data, the proposed rules have not specified whether the weighted average or simple average should be used for the application of the multiple-year data (in the manner as explained in the rules relating to the range). A specific provision in the rules on the use of the weighted average in the case of multiple-year data would remove the existing ambiguity in the manner of computation.

It is worthy to note that the proposed rules provide for the use of multiple-year data of the comparable companies, but are silent on whether the tested party’s data to be used shall be only in relation to the financial year in which the transaction has taken place, or a corresponding multiple-year period. In such a scenario, the provisions are open to interpretation by both the taxpayer and the revenue authorities, which again can lead to protracted litigation.

The introduction of the ‘range concept’ and ‘use of multiple year data’ can clearly be regarded as a move in the right direction by the CBDT to align the Indian TP provisions in line with globally accepted practices. However, its restrictive application to the arm’s length determination under the margin based methods, the requirement of at least nine comparable companies by all taxpayers and the restrictive percentiles of 40%-60% for the construction of the range could diminish the success of these newly introduced concepts. We hope the CBDT will resolve these issues before issuing the final rules.

(Views expressed are personal)


Copyright © Ltd. All rights reserved. Reproduction of news articles, photos, videos or any other content in whole or in part in any form or medium without express written permission of is prohibited.