www.pwc.in
Recommendations on the Draft Scheme of
the proposed rules for computation of ALP
- “range” concept and “multiple year data”
May 2015
Table of contents
I.
Our sincere appreciation of the Government’s initiatives
3
II.
Our recommendations – A snapshot
4
III.
Our recommendations – In detail
8
IV.
Specific clarification sought
18
V.
A concluding thought
18
May 2015
Page | 2
I.
Our sincere appreciation of the Government’s initiatives
At the outset, we sincerely appreciate and applaud the steps being taken by the Indian Government in the direct tax arena of transfer pricing. In
this regard, apart from the recent measures taken in respect of APA roll backs and MAP resolution framework, the introduction of the range
concept and use of multiple year data in the Finance Budget 2014, have, in principle, been very well received by the industry. These measures
evidence the intent of the Government to meet its stated objectives of reducing TP litigation, resolving TP disputes, facilitating ease of doing
business, and bringing Indian regulations in line with internationally followed practices.
Further, the approach adopted by the Government of seeking stakeholder and general public comments on the Draft Scheme with respect to use of
“range” concept and “multiple year data” (hereinafter referred to as ‘the draft scheme’) is clearly inclusive and collaborative. It is also in line with
best practices followed globally, and will certainly help in building trust and enhancing taxpayer confidence.
However, it is important that the draft scheme translates into actually meeting the stated objectives of the Government “on-ground”. It is in this
light that we make our humble recommendations with respect to the draft scheme.
(This space has been intentionally left blank)
May 2015
Page | 3
II. Our recommendations – A snapshot
We first provide below a bird’s eye view of our recommendations in light of the stated objectives of the Government. Our recommendations are
intended to resolve practical and implementation difficulties that may likely arise when the draft scheme is in fact implemented:
Objectives of the
Government ----->
Reduced
litigation
Proposed rules
Taxpayer
confidence,
certainty
and ease of
doing
business
Alignment
with global
best/
internationally
followed
practices
Alignment with
Alternate
Dispute
Resolution
Mechanisms
(APA, MAP etc.)
Our recommendations
Are Government’s objectives being met?
A minimum of 9
entities are required to
be selected as
comparable entities,
based on similarity of
FAR

Range defined as 40th
to 60th percentile of
the data set or series




The requirement for a minimum number of
comparables should be done away with.
As an alternative to specifying the minimum
number of comparables, the rules could be
modified to provide for broadening or
expansion of the set of comparables to close or
related industries, and the range should be
used on the comparable set emanating
therefrom. In such a situation, use of range
could in fact be preferred in all circumstances,
rather than resorting to arithmetic mean.



The proposed range should be broadened to an
inter-quartile range in line with internationally
followed practices.
(This space has been intentionally left blank)
May 2015
Page | 4
Objectives of the
Government ----->
Proposed rules
“The data points
lying within the
40th to 60th
percentile of the data
set of series would
constitute the range.”
(emphasis supplied in
bold italic text)
Reduced
litigation

Taxpayer
confidence,
certainty
and ease of
doing
business
Alignment
with global
best/
internationally
followed
practices
Alignment with
Alternate
Dispute
Resolution
Mechanisms
(APA, MAP etc.)
Are Government’s objectives being met?
Not


applicable
Formula for
computation of range
– currently not
specified in the rules


Not applicable
Not applicable
The range concept and
use of multiple year
data shall be
applicable only in
cases where the most
appropriate method is
TNMM/ RPM/ CPM.

Partly

Partly
Our recommendations
The phraseology “data points lying within”
could create ambiguity or in some situations
may also render the use of range redundant.
Accordingly, without prejudice to our
recommendation of adopting inter-quartile
range instead of the proposed percentile based
range, we recommend that the language of para
A(ii)(d) of the draft scheme be modified as
follows:
“The 40th to 60th percentile of the data set or
series would constitute the range.”
Since multiple formulae for computing range
could exist, the simple and non-controversial
excel-based formula for computation of the
lower end and upper end of the range should be
specifically provided for in rules as the formula
to be applied.
Extend the use of range concept to all methods.
Further, extend the use of ‘multiple year data’
to Profit Split Method. This would be relevant
for the purposes of ascertaining routine returns
under the residuary profit split method, based
on comparable margins typically determined by
applying a methodology akin to TNMM, RPM
or CPM, for which use of multiple year data and
the range concept have been provided in the
draft scheme.
May 2015
Page | 5
Objectives of the
Government ----->
Reduced
litigation
Proposed rules
In case of nonavailability of data of
comparables for 3
years, the use of data
of two out of relevant
three years shall be
permitted, subject to
exceptions.
Taxpayer
confidence,
certainty
and ease of
doing
business
Alignment
with global
best/
internationally
followed
practices
Alignment with
Alternate
Dispute
Resolution
Mechanisms
(APA, MAP etc.)
Our recommendations
Are Government’s objectives being met?
Exception 1: Data for
current year may not
be available on the
databases at the time
of filing of returns
Partly
Exception 2: A
comparable may fail to
clear a quantitative
filter in any one out of
the three years
Partly
Exception 2:
“A comparable may
fail to clear a
quantitative filter in
any one out of the
three years.”
(emphasis supplied in
bold italic text)















Partly





Partly

























Partly
Partly










Partly
Partly












This exception should be extended to prior two
years data as well. It is therefore recommended
that the language of the first exception be
modified as follows:
“Unavailability of data for any one of the three
years at the time of filing return of income.”
A similar exception should be extended to a
comparable not passing a qualitative filter.
Accordingly, the following should be added to
the specific exceptions:
“A comparable may fail to clear a qualitative
filter in any one out of the three years.”
The language of this exception makes it a
stringent exception and may lead to rejection of
greater number of comparables. It is therefore
recommended that the language of this
exception be modified as follows:
“A comparable may fail to clear a quantitative
filter computed on the basis of average data of
at least two out of the three years.”
May 2015
Page | 6
Objectives of the
Government ----->
Reduced
litigation
Proposed rules
Taxpayer
confidence,
certainty
and ease of
doing
business
Alignment
with global
best/
internationally
followed
practices
Alignment with
Alternate
Dispute
Resolution
Mechanisms
(APA, MAP etc.)
Our recommendations
Are Government’s objectives being met?
Data of the current
year can be used
during the transfer
pricing audit by both
the taxpayer and the
department if it
becomes available at
the time of audit.

3-year data of the 9
comparable entities
(or more) would be
considered
Partly



Partly


It is recommended that this provision be done
away with, as it is not in line with the
contemporaneous documentation
requirements under the Indian TP regulations
[Rule 10D(4) of Income-tax Rules, 1962].
Currently the draft scheme contemplates use of
multiple year data only of comparables.
To ensure a like-to-like comparison, it is
recommended that the use of multiple year
data be extended for the tested party as well.
This is followed internationally in certain
countries.
Key:
 No
  Yes
(This space has been intentionally left blank)
May 2015
Page | 7
III. Our recommendations – In detail
1. Number of comparables
Issue: The draft scheme restricts the applicability of the range concept only when there is a minimum of 9 comparable entities.
Our recommendation: The requirement of a minimum number of comparables should be done away with.
As an alternative to specifying the minimum number of comparables, the Rules could be modified to provide for broadening or expansion of the set
of comparables to close or related industries, and the range should be used on the comparable set emanating therefrom. An example of a country
which adopts such an approach is Russia. In such a situation, use of range could in fact be preferred in all circumstances, rather than resorting to
arithmetic mean.
Underlying rationale for our recommendation:
The underlying rationale for our recommendation of doing away with the requirement of a minimum number of comparables is as follows:

In our experience during TP audits and even for documentation / pricing purposes, for most industries/ nature of transactions, except in the
case of Information Technology (IT) service companies or IT enabled services companies, 9 comparables have been difficult to get (kindly refer
to Annexure A for some of the industries/ nature of transactions for which the number of comparables is generally less than 9).
Moreover, the final number of comparables is a market oriented economic outcome beyond the control of taxpayers and Revenue authorities,
as it is dependent upon the comparability at a FAR level (with the taxpayer) and the number of industry players whose data is available in
public domain. Further, as for information available in public domain (including databases of companies) – the same is not yet at such a stage
of evolution so as to enable yield of 9 or more comparables across industries.
The minimum requirement of 9 comparables would thus in effect limit the applicability of ‘range’ concept to only a small number of taxpayers.

Specifying the number of comparables could lead to a situation where the taxpayer, for example, uses 9 comparables for the purpose of price
setting and its TP documentation and the assessing officer reduces them to say 7 at the time of TP audit – then, as per the draft scheme, the
taxpayer would be unable to apply the “range” and will have to move back to “arithmetic mean” at the time of TP audit. The reverse of this is
also possible wherein the taxpayer uses less than 9 comparables (and therefore the arithmetic mean) to set its prices and also for TP
documentation, but at the time of TP audit the assessing officer adds comparables such that they exceed 9, and hence applies the range.
May 2015
Page | 8
Such situations could lead to a fair amount of ambiguity, uncertainty and reconciliation difficulties for most taxpayers and render their TP
documentation and price setting redundant.

Jurisdictions which use inter-quartile range may provide a limit of 4 comparables (as inter-quartile range is typically best applied to a
minimum of 4 comparables). To the best of our knowledge, however, no jurisdictions specify a requirement for the number of maximum or
minimum comparables. The Indian rule of specifying a minimum number of 9 comparables would thus be unique and a deviation from
internationally followed practices. Moreover, specifying the number of comparables would leave room for convenient manoeuvrability.

Since the restriction of minimum number of 9 comparables would be a deviation from what is generally followed by other jurisdictions, it may
not find acceptance by Competent Authorities of other countries during a bilateral advance pricing arrangement (APA), multilateral APA or a
mutual agreement procedure (MAP) negotiation. Given the Government’s current focus on dispute resolution through MAPs and APAs – this
could prove detrimental.
2. 40th to 60th percentile
Issue: The range of 40th to 60th percentile is a very narrow range as most of the 9 comparables (or more) would become redundant when using
this range as only 20% of the set of finally selected comparables will eventually fall within the range.
Our recommendation: The proposed range should be broadened to an inter-quartile range (IQR).
Underlying rationale for our recommendation:

Adoption of the IQR would be in line with globally followed practices. There are a host of countries which follow the IQR either because it is
part of their respective transfer pricing legislations, or simply as a matter of practice. Annexure B provides a list of all such countries and a
brief description of the approach adopted by them with respect to IQR. To the best of our knowledge, however, no jurisdictions specify a
percentile range as in proposed in the draft scheme. The Indian rule would thus be unique and a deviation from internationally followed
practices.

Since the use of a percentile range would be a deviation from what is generally followed by other jurisdictions, it may not find acceptance by
Competent Authorities of other countries during a bilateral advance pricing arrangement (APA), multilateral APA or a mutual agreement
procedure (MAP) negotiation. Given the Government’s current focus on dispute resolution through MAPs and APAs – this could prove
detrimental.
May 2015
Page | 9

The concept of range was introduced to reduce TP litigation. However, the proposition to use a narrow 40 th to 60th percentile is unlikely to
achieve this objective as is evident from some examples below:
Situation 1:
Comparables
Weighted average
(Operating profit/
Total cost)
Company 1
2.72%
Company 2
7.89%
Company 3
8.82%
Company 4
2.87%
Company 5
11.33%
Company 6
9.22%
Company 7
11.23%
Company 8
12.67%
Company 9
5.45%
Arithmetic Mean
8.02%
PLI of tested party
7.12%
Tolerance band +3%
10.33%
Tolerance band -3%
3.91%
Median
8.61%
40th Percentile
8.08%
60th Percentile
9.14%
Upper quartile (of IQR)
11.23%
Lower quartile (of IQR)
5.45%
Evident from this example are the following observations:
a) Applying the arithmetic mean along with the tolerance
band, the taxpayer’s transfer prices are at arm’s
length.
b) However, if the 40th to 60th percentile range is
applied, the taxpayer is out of range and an
adjustment would be warranted. An adjustment
would, however, not be required if an inter-quartile
range is used.
May 2015
Page | 10
Situation 2:
Let’s take another situation where applying the arithmetic mean along with the tolerance band, the taxpayer’s transfer prices are not at arm’s
length.
Comparables
Weighted average
(Operating profit/
Total cost)
Company 1
2.72%
Company 2
7.89%
Company 3
8.82%
Company 4
2.87%
Company 5
11.33%
Company 6
9.22%
Company 7
11.23%
Company 8
12.67%
Company 9
3.45%
Arithmetic Mean
7.80%
PLI of tested party
4.20%
Tolerance band +3%
7.32%
Tolerance band -3%
1.07%
Median
8.61%
40th Percentile
8.08%
60th Percentile
Evident from this example are the following observations:
a) Applying the arithmetic mean along with the tolerance
band, the taxpayer’s transfer prices are not at arm’s
length, and an adjustment would be warranted.
b) Even if the 40th to 60th percentile range is applied,
the taxpayer is out of range and an adjustment would
be warranted. However, the adjustment hereby would
be greater in quantum as compared to an adjustment
made using the arithmetic mean.
c) An adjustment would, however, not be required at all
if an inter-quartile range is used.
9.14%
Upper quartile (of IQR)
11.23%
Lower quartile (of IQR)
3.45%
May 2015
Page | 11
It is apparent that in the above situations, the use of 40 th to 60th percentile range is in fact detrimental to the taxpayer as against use of interquartile range or even the arithmetic mean (along with the tolerance band). Notably, in both situations the tolerance band provides a much larger
range which in fact gets hugely restricted under the 40 th to 60th percentile range – and this could certainly not have been the intent of the
Government.
The above are simply illustrations – but there could be several such real situations where TP adjustments would now be made under the new rules
even when they were not being made under the erstwhile regime or being made but to a lower extent. Therefore, instead of curbing TP disputes –
the proposed rules may in fact proliferate greater litigation if such a narrow range of 40 th to 60th percentile is used. This would not be so, however,
if under the new rules inter-quartile range is allowed.
Accordingly, the liberal step taken by the Government to adopt a ‘range’ concept must be complimented by use of a wider range, lest it nullifies the
very underlying objective of the Government.
Apart from the mathematics as discussed above, from a commercial standpoint as well such a limited set of comparables would not be truly
representative of the margins prevailing in the industry. In principle, comparison with a wider set of comparables would always enable a better
evaluation of the taxpayer’s performance vis-à-vis the performance of the industry at large.
3. Mechanics of the 40th to 60th percentile
Issue: The draft scheme [in para A(ii)(d)] prescribes that the “The data points lying within the 40th to 60th percentile of the data set of series
would constitute the range”. The phraseology “data points lying within the 40th to 60th percentile” could create ambiguity or in some situations
may also render the use of range redundant.
Our recommendation: Without prejudice to our recommendation of adopting IQR instead of the proposed percentile based range, we
recommend that the language of para A(ii)(d) should be modified to say “ The 40th to 60th percentile of the data set or series would constitute the
range.”.
(This space has been intentionally left blank)
May 2015
Page | 12
Underlying rationale for our recommendation:
The ambiguity that may arise and the possibility of use of range becoming redundant can be explained by means of the example below:
Comparables
Scenario 1
Scenario 2
Company 1
-1.00%
-1.00%
Company 2
1.00%
1.00%
Company 3
4.00%
4.00%
Company 4
7.00%
7.00%
Company 5
7.10%
7.10%
Company 6
7.50%
8.00%
Company 7
8.00%
9.00%
Company 8
9.00%
9.00%
Company 9
9.00%
10.00%
Company 10
10.00%
-
Arithmetic Mean
6.16%
6.01%
40th
Percentile
7.06%
7.02%
60th
Percentile
7.70%
7.82%
7.30%
7.10%
Median
Under Scenario 1 above, the range as per the current language of the draft scheme would be 7.10% to 7.50% (being data points lying within 40th to
60th percentile), and not 7.06% to 7.70% (being the 40th and 60th percentile, respectively). Similarly, under Scenario 2, the range concept will fail
as there is only one data point that lies within the 40th to 60th percentile. Accordingly, the language of the draft scheme would need to be
modified in order to remove this ambiguity. The modification recommended above is also in line with generally followed international practices.
(This space has been intentionally left blank)
May 2015
Page | 13
4. Formula for computation of range
Issue: The draft scheme does not provide the mathematical formula to be used for computation of the lower end and upper end of the range.
Multiple formulae to compute range could exist. Therefore, the lack of any specific provision in this regard in the rules could create confusion.
Our recommendation: The excel-based formula for computation of the lower end and upper end of the range should be specifically provided
for in rules as the formula to be applied.
Underlying rationale for our recommendation: The formula specified should be simple so as to avoid any controversies or complications.
Since this is, in essence, purely a mechanical aspect of the rules, it would be best to prescribe use of the simple and non-controversial excel-based
formula for computing the upper and lower end of the range.
5. Restricted applicability to methods
Issue: The draft scheme proposes to allow application of range concept and use of multiple year data only in case determination of ALP is by
Transactional Net Margin Method (TNMM), Resale Price Method (RPM) or Cost Plus Method (CPM).
Our recommendation: Extend the use of ‘range’ concept to all methods.
Further, extend the use of ‘multiple year data’ to Profit Split Method. This would be relevant for the purposes of ascertaining routine returns
attributable to an enterprise under the residuary profit split method, based on comparable margins.
Underlying rationale for our recommendation:
TNMM, RPM or CPM are profit based methods as against the Comparable Uncontrolled Price (CUP) Method and the Other Method which are
price based methods. Achieving a specific price benchmark may be relatively more difficult than achieving a profit benchmark, and thus greater
flexibility may be required in case of the former. Application of the range concept would provide such flexibility because the tested party’s price
would need to fall within a range rather than be a specific price (arithmetic mean). In case of prices, even the tolerance band (applied along with
arithmetic mean) would be a quite limited.
Examples of transactions, wherein application of the range concept would prove beneficial, are royalties, interest payments, guarantee fees etc.
The flexibility provided by application of range to prices will go a long way in curbing litigation in respect of such transactions and bringing
consistency within the application of methods prescribed in the Indian TP regulations.
May 2015
Page | 14
As regards application of range concept and use of multiple year data being extended to Profit Split Method (PSM) – this would be needed as PSM
under the residuary method requires that the parties to the transaction be first remunerated for the ‘routine functions’ performed by them. The
routine returns, in such cases, are typically determined by applying a methodology akin to TNMM, RPM or CPM, for which use of multiple year
data and the range concept have been provided in the draft scheme.
It is also important to note that to the best of our knowledge, other jurisdictions do not restrict the applicability of range concept to any particular
method/s.
6. Exceptions to the use of three year data
Issue 1: The draft scheme specifies exceptions to the use of three year data. One of the exceptions is when data for current year is unavailable at
the time of filing return of income.
Our recommendation: It is recommended that this exception be extended to prior two years data as well. It is therefore recommended that the
first exception be modified as follows: “Unavailability of data for any one of the three years at the time of filing return of income. ”
Underlying rationale for our recommendation:
There could be situations wherein a comparable has data for the current year but does not have data for one of the prior two years (regardless of
the reason for the same). The comparable could otherwise be a valid comparable for the current year and the remaining one year. If the exception
relating to “unavailability of data” is not extended to the prior two years as well, then there is a possibility of losing out on good comparables.
Therefore, the language of the first exception should be modified as recommended above.
Issue 2: The draft scheme specifies exceptions to the use of three year data. One of the exceptions is when a comparable fails to clear a
quantitative filter in any one out of the three years. There is, however, no similar exception provided for a comparable not passing any of the
qualitative filters.
Our recommendation: The following be added to the specific exceptions: “A comparable may fail to clear a qualitative filter in any one out of
the three years.”
Underlying rationale for our recommendation:
There could be several situations wherein a comparable does not clear a qualitative filter for one out of the three years, for instance, on account of a
change in its functional profile for that particular year or the non-availability of segmented profit and loss account, etc. The comparable could
otherwise be a valid comparable for the remaining two years. If the exception to use two year data is not extended to such cases, there is a
May 2015
Page | 15
possibility of losing out on good comparables. Therefore, even if the comparable does not clear a qualitative filter in one out of the three years, use
of its financial data of the remaining two years should be allowed.
Issue 3: The draft scheme specifies exceptions to the use of three year data. One of the exceptions is when “ A comparable may fail to clear a
quantitative filter in any one out of the three years”. The language of this exception makes it a stringent exception and may lead to rejection of
greater number of comparables.
Our recommendation: The language of this exception should be modified as follows: “A comparable may fail to clear a quantitative filter
computed on the basis of average data of at least two out of the three years. ”
Underlying rationale for our recommendation:
A reading of the provision as it stands implies that each company would need to be tested separately on a year on year basis, in order to check if it
clears the relevant quantitative filter. This may lead to rejection of greater number of comparables as is demonstrated in the example below:
Sales (in INR crores)
Company
2013
2014
2015
Average
3.75
3.51
Company 1
2.50
4.27
Company 2
1.76
0.58
1.17
Company 3
3.50
2.37
2.94
If, for instance, a quantitative filter of Sales >= INR 1 crore is applied to the average data, then all the companies would be accepted. However, if
the filter is applied on an individual year wise basis (as is contemplated in the draft scheme), Company 2 would be rejected.
The rules should enable acceptance of a greater number of comparables so as to improve the overall comparability. Moreover, the use of average
data would in essence also be in line with the proposition in the draft scheme to compute weighted average of three year data to construct a set of
comparables, and hence internally consistent.
7. Use of current year data during TP audit
Issue: The draft scheme allows the use of current year data if it becomes available at the time of TP audit.
Our recommendation: It is recommended that this provision be specifically done away with.
May 2015
Page | 16
Underlying rationale for our recommendation:

Use of financial data of comparables which was not available in the public domain on or before the specified date (i.e., due date for filing of the
income tax return for the relevant tax year), but which becomes available subsequently by the time of audit, is not in line with the
contemporaneous documentation requirements under the Indian TP regulations [Rule 10D(4) of Income-tax Rules, 1962].

Allowing use of current year data which only subsequently becomes available at the time of TP audit, would revive the perennial concern of
taxpayers’ TP documentation becoming a nullity, i.e., being rendered otiose.
8. Use of multiple year data for tested party
Issue: Currently the draft scheme contemplates use of multiple year data only of comparables.
Our recommendation: To ensure a like-to-like comparison, it is recommended that the use of multiple year data be extended for the tested
party as well. This is followed internationally in certain countries, for example, China, Denmark, Netherlands, and USA.
Underlying rationale for our recommendation:
There could be situations wherein the industry had a bumper year for a few years, after which the performance of the industry normalised. Take
for example, in the table below, year 1 and year 2 were bumper years, whereas in the current year, the performance of the industry normalized. At
the time of filing return of income, comparable data was available only for year 1 and year 2. In the below example, if the tested party results for
the current year are compared to that of comparables for prior two years, then the tested party would on the face of it not be at arm’s length.
However, owing to the business cycle which the industry went through that would not be a fair conclusion because if the results of the tested party
for all three years are compared to the results of the comparables for the same three years, then the tested party would be at arm’s length. Just like
use of multiple year data evens out the impact of business cycles for comparables – it does so for the tested party as well.
Comparables
Company 1
Year 1
35%
Year 2
Current
year
42%
14%
28%
49%
11%
30%
40%
9%
45%
12%
Company 2
Company 3
Tested party
32%
May 2015
Page | 17
IV. Specific clarification sought
The draft scheme in para A(iii) provides that “There shall not be two different data sets – one for testing and one for making adjustments.” It is not
clear from this provision as to under what situations would there be two different data sets (i.e., one for testing and one for making adjustments).
We request the Government to clarify this provision to avoid ambiguity because the provision as it currently stands leaves scope for multiple
interpretations.
V. A concluding thought
Our recommendations proposed above are intended to align the draft scheme with the stated objectives of the Government.
The Indian Income-tax Act and the Rules thereto, which are aligned with global best/ internationally followed practices, and which provide
certainty to taxpayers and facilitate ease of doing business, would not only achieve the objectives of the Government from an income-tax
standpoint, but may also serve as a worthy precedent and guiding force for other regulations containing provisions in respect of related party
transactions. Such other regulations include the Companies Act, 2013 (sections 177 and 188), SEBI’s Clause 49 of the Equity Listing Agreement,
and the Customs regulations. Provisions relating to arm’s length pricing in these other regulations are either archaic or limited, and therefore the
principles emanating from the Income-tax Act and Rules thereto may provide useful guidance and direction for companies while complying with
these other regulations.
Going forward, harmonization of these other regulations with the Income-tax Act and Rules thereto would become critical to ensure simplicity of
compliance, certainty of outcomes, and overall ease of doing business.
(This space has been intentionally left blank)
May 2015
Page | 18
Annexure A – Number of comparables in different industries/ transactions
Industry / Nature of transaction
Number of comparables
based on
experience / trends
Clinical Research Services
4
Contract Research & Development Services
7
Business Support Services
6
Business Services
5
Contract Manufacturing Chemicals & Pharmaceuticals
3
Courier Services
4
Design Engineering Services
5
FMCG Distribution activity
5
Logistics Services
7
Marketing Support Services
3
Marketing of financial products
7
API- Manufacturing
5
Pharmaceutical Distribution Activity
4
Software Distribution Activity
6
Toll Manufacturing Activity
5
May 2015
Page | 19
Annexure B – International precedence with respect to use of IQR
Sr
no
1
Country
Australia
Approach adopted with respect to IQR
The ATO has not specified a preferred range to be used when setting or reviewing transfer prices, but will
generally refer to the interquartile range.
In cases where there are a limited number of observations available in the set of comparables, the full range
of results may be considered in determining the arm's length range/price.
2
China
The specified transfer pricing methods for intercompany transactions under CIT Implementation
Regulations and the Measures are similar to those of the OECD Guidelines. Although the Chinese
regulations do not directly specify the use of any particular type of range, in practice the interquartile range
is generally accepted by the Chinese tax authorities as proxy for the arm’s length price.
If the profit level is below the median value of a range of profit levels of comparable enterprises, tax
authorities generally, in principle, make an adjustment based on a profit level not lower than the median
value.
3
France
The transfer pricing methods listed in the French legislation are generally in line with the OECD
Guidelines. In France, the use of an arm’s length range is not explicitly specified in the legislation, but
common practice is to use the interquartile range.
4
Germany
In the case of restricted comparability, which is the most likely case in practice, the interquartile rang e is
preferred.
In the case of full information and if third-party prices are fully comparable, the taxpayer may use the full
range. In cases where only limited reliable data is used, the range must be narrowed, e.g. though the use of
an interquartile range. If the prices are outside the range of admissible prices, an adjustment to the median
shall be made.
May 2015
Page | 20
Sr
No.
5
Country
Netherlands
Approach adopted with respect to IQR
The transfer pricing methods listed in the Dutch legislation are generally in line with the OECD Guidelines.
Dutch legislation does not specify a range; however, a Decree issued by the Ministry of Finance gives the
example of an interquartile range while explaining OECD’s definition of an arm’s length range.
As per the above mentioned Decree, where comparables are accurate, then the range could be defined as
ranging from maximum to minimum, i.e., all the values would fall in the range. However, if the
comparables are less accurate, then statistical methods may be used to increase the reliability of the data,
for example, use of an inter-quartile range.
6
Singapore
Singapore regulations are consistent with those of the OECD Guidelines. Although the Singapore
regulations do not specify the use of a specific arm’s length range, the interquartile range is generally
accepted by the tax authorities in practice.
Singapore transfer pricing guidelines recommend the use of an arm’s length range, and the inter-quartile
range is widely used. It is generally accepted that any point in the inter-quartile range (and not necessarily
always the median) would be arm’s length.
7
Sweden
The Swedish specified transfer pricing methods are generally consistent with the OECD Guidelines. The use
of a range is not specified in the legislation, but the use of an interquartile range is common practice.
Corporations are advised to update their benchmarking studies every two years assuming that no major
changes took place in the organisation or within the industry.
May 2015
Page | 21
Sr
No.
8
Country
United States
Approach adopted with respect to IQR
The US regulations allow use of interquartile range which is the middle 50% of the observations. IRS
practice is to utilize an interquartile range in virtually every case.
The section 482 regulations require that the actual results of controlled transactions be within the arm’s
length range on an after-the-fact basis; however, there is no requirement that the actual results hit any
specific point within the range, such as the median or the mean.
9
Denmark
All results within the range that meet the comparability standards in principle indicate an arm’s length
benchmark. Comparables are often, however, subject to elements of uncertainty due to lack of information.
Statistical tools, including an interquartile range, are applied to minimize comparability uncertainties.
10
Italy
Even if there is no formal guidance in this regard, based on the practice of the tax authorities,
internationally adopted criteria are accepted (the so-called interquartile range).
Also, general practice is to make an adjustment to the median.
11
Korea
Interquartile ranges are commonly used in audits, as well as APAs and MAPs. On 15 June 2004, the NTS
Basic Guidelines adopted the concept of an interquartile range, which is calculated as the middle half of the
full range. No clear rule on using loss companies. Various other statistical tools may be acceptable to the tax
authorities, depending on the circumstances.
12
South Africa
OECD is followed – it is accepted that the adjustment should reflect the point in the range that best
accounts for the facts and circumstances of the controlled transaction. However, in the absence of
persuasive evidence for the selection of a particular point in the range, the mid-point in the range may be
selected.
May 2015
Page | 22
Sr
No.
Country
Approach adopted with respect to IQR
13
Finland
As a starting point all points within the range may be regarded as at arm’s length. In cases where the result
of the tested party is not within the arm’s length range, transfer prices would be adjusted to the median of
the benchmark study. Although, the authorities, in their guidance, explain how the interquartile range is
calculated – however, the guidance does not explicitly state whether the whole range or preferably only the
interquartile range should be used.
14
Malaysia
Malaysian Transfer Pricing Guidelines 2012 state that an arm’s length range refers to a range of figures that
are acceptable in establishing the arm’s length nature of a controlled transaction. In practice, the
interquartile range is used for a given sample of comparables as a starting point. The median result for such
sample is then compared against that earned by the taxpayer.
15
Spain
Spanish tax legislation does not specify whether full range, interquartile range or other range is preferred.
In practice it depends on the specific circumstances, for example, if there are 5 comparable companies
which are highly comparable, the full range can be accepted.
16
Thailand
There is no guideline on the statistical tool that should be used to determine an arm’s length range. In
practice, the Revenue officers accept an inter-quartile range. As a result, extreme results are not taken into
consideration.
17
United Kingdom
There is no rule in the United Kingdom to determine what is or is not an arm’s length range. Where there
are several good comparables, the full range may be used in accordance with the stipulations of the OECD
Guidelines on the use of ranges. Where there are a number of reasonable (rather than good) comparables,
which is more usually the case, it is common to use an interquartile range and attention may be paid to the
median or, less commonly, the mean result. No other statistical tools are in common use for evaluating
company results.
May 2015
Page | 23
About PwC
PwC helps organisations and individuals create the value they’re looking for. We’re a network of firms in 157 countries with more
than 195,000 people who are committed to delivering quality in Assurance, Tax and Advisory services.
PwC India refers to the network of PwC firms in India, having offices in: Ahmedabad, Bangalore, Chennai, Delhi NCR, Hyderabad,
Kolkata, Mumbai and Pune. For more information about PwC India’s service offerings, please visit www.pwc.in.
*PwC refers to PwC India and may sometimes refer to the PwC network. Each member firm is a separate legal entity. Please see www.pwc.com/structure for further details. Tell us
what matters to you and find out more by visiting us at www.pwc.in
For private circulation only
This publication has been prepared for general guidance on matters of interest only, and does not constitute professional advice. You should not act upon the information contained in this publication without
obtaining specific professional advice. No representation or warranty (express or implied) is given as to the accuracy or completeness of the information contained in this publication, and, to the extent permitted by
law, PwCPL, its members, employees and agents accept no liability, and disclaim all responsibility, for the consequences of you or anyone else acting, or refraining to act, in reliance on the information contained in
this publication or for any decision based on it. Without prior permission of PwCPL, this publication may not be quoted in whole or in part or otherwise referred to in any documents.
© 2015 PricewaterhouseCoopers Private Limited. All rights reserved. In this document, “PwC” refers to PricewaterhouseCoopers Private Limited (a limited liability company in India having Corporate Identity
Number or CIN : U74140WB1983PTC036093), which is a member firm of PricewaterhouseCoopers International Limited (PwCIL), each member firm of which is a separate legal entity.
May 2015
Page | 24