The Central Board of Direct Taxes (CBDT) has recently issued the much awaited draft rules for the use of multiple year dataand application of the concept of range,for the purposes of computing of arm's length price (ALP), post the announcement made by the Finance Minister, in the budget speech for 2014, for the introduction ofthe said concepts in the Indian transfer pricing regulations (TP), in place of single year data and arithmetic mean, which have been in vogue since the inception of the formal TP regulations in India; and have fuelled the majority of TP adjustments in India, particularly in the context of comparability analyses. The usage of single year data and arithmetic mean significantly deviated from the guidelines of the OECD; and also the best practices followed by large number of countries having TP regulations, falling under the brackets of both developed and developing economies.
The concept of range as introduced in the draft rules is proposed to be applicable only while determining ALP by application of Transactional Net Margin Method (TNMM), Cost Plus Method (CPM) and Resale Price Method (RPM). The CBDT proposes that a minimum of nine comparables would be required in the data set for application of the concept of range. The data points lying between the 40th and 60th percentile of the data set shall constitute the range. If the margin of the tested party falls outside the data range, TP adjustments shall be made with reference to the median of the range. However, in absence of the minimum of nine comparables, the concept of range would not apply; and resort would be had tothe existing provision of arithmetic mean, along with the benefit of the tolerance band of 3% (1% in the case of wholesale trading companies). Incidentally, the OECD and UN TP guidelines; and also tax authorities across the world, namely of both developing and developed countries,
adopt the classical inter-quartile range, namely the range between the 25th and 75th percentiles; and the concept of range proposed by the CBDT, as above, is quite unique; and it is not sure as to whether implementing the same would absolve the objective of mitigating TP litigation, as we shall try and explain through empirical data, later on in this article.
For the use of multiple year data in determination of ALP, the rules provide for use of data for three years including the data for the current year in which the transaction was undertaken. However, use of data for two out of three years shall be permitted in cases where- (a) data for current year is not available in the public domain at the time of preparation of transfer pricing documentation; (b) any comparable company has commenced the operation only in the last two years; or (c) any comparable company fails to clear a quantitative filter in any one out of three years. However, if current year data is not available at the time of preparation of TP documentation, but becomes available during the TP audit, it can be used by both the taxpayers and Revenue Authorities.
While using multiple year data, weighted average of the data for three years for each company shall be used to construct the data set.
In order to assess the results of the new proposed concept on the taxpayers, we have applied the same to standard sets of comparables, which were adopted by Revenue Authorities for captive contract service providers in the fields of IT and ITeS, in the most recently concluded audit cycle. We have taken a ball-park margin of 15%, being the mark up considered to be generally applied by taxpayers falling in such sectors. Applying the tolerance band of 3% in the context of arithmetic mean, a taxpayer of the above genre, having a mark-up of 15%, would be at arm's length, if the arithmetic mean of the comparables is 18.45% or below.Based on the experience of the recently concluded audit cycle, the arm's length margin adopted by the Revenue Authorities for IT and ITeS sectors were 24.14 and 27.19% respectively, where each of the sets had comparable companies in excess of nine.The results of the statistical analyses carried out on the above sets are set out in the chart, as
On a perusal of the aforesaid results, it would be evident that while a mark-up of 15% falls within the classical inter-quartile ranges of both the sets, so as to qualify as ALP, the same falls outside the ranges in the manner computed as per the proposals of CBDT, namely the range between the 40th and 60th percentiles of the data sets, so as to suffer TP adjustments with reference to the medians, which are merely lower than the arithmetic means.
It is not the case that captive service providers in the fields of IT and ITeS, operating with a mark-up of 15% on cost should never suffer any TP adjustment, irrespective of the compositions of the benchmarking sets, however, when it is palpably clear that by applying the classical concept of inter-quartile range, which is advocated by the OECD and UN; and also by Revenue Authorities across the world, such mark-up of 15% clearly qualifies as the ALP for both the sets of IT and ITeS, then it is most unfortunate for the same mark-up facing TP adjustments with references to a convoluted or distorted concept of range, ad proposed by the CBDT, i.e. between the 40th and 60th percentiles.
It is recommended that:
- The concept of range should be made consistent with that followed by Tax Authorities across the world; and also professed by the OECD. It would be much easier for MNCs to comply with the Indian TP regulations, if similar standards are adopted by both countries involved in a transaction, thus facilitating easier negotiation and resolution of bilateral APA and MAP applications.
- Adoption of the classical inter-quartile range would also reduce the minimum number of comparables required for reliable application of the concept of arm's length range, as inter-quartile range can be computed with a minimum number of four comparables.
The rules proposed by the CBDT provide for a minimum number of nine comparables for application of the concept of range. While it may not be a big challenge to find the required number of comparables for IT, ITeS, etc sectors, it might make the concept redundant for certain sectors due to paucity of data in the public domain. Further, the requirement of having a minimum number of nine comparables for applying the concept of range with reference to the 40th and 60th percentiles, might actually entice Revenue Officers to deliberately reject certain comparables, initially selected by the taxpayer, so as to push the case of the taxpayer towards the rigours of arithmetic mean. Thus, it is imperative for the CBDT to introduce the classical concept of inter-quartile range, in case the CBDT seriously wishes to mitigate TP litigation around comparability analyses; and align the TP regulations of India to match the best in the world.
- While the proposed rules have proposed for the application of concept of range for RPM, CPM and TNMM, it has not been extended to economic analysis undertaken adopting Comparable Uncontrolled Price (CUP) method. It needs to be appreciated that data points may also be available while benchmarking transactions by application of CUP method; and in such a scenario too, it would be essential to discard the outliers for better comparability analyses. Hence, it is recommended that the application of concept of range should be extended to CUP method as well.
Overall, the introduction of concepts of range and multiple year data is a welcome step. The concept of range would provide a more accurate result for ALP, as extreme results or outliers would be left out of the range, which was not the case with the use of arithmetic mean. Similarly, the usage of multiple years' data shall help in improving the overall exercise of comparability analyses. A comparability analysis using multiple years' data provides useful insights in understanding long term arrangements, business and product cycles, identifying results indicating variance in comparability characteristics of controlled transactions, etc. The move is a step towards alignment of the Indian TP regulations with global best practices, subject to incorporating the suggestions or recommendations, as above.
This article has also been co-authored by Devendra Gulati (Senior Manager).