Philips Software Centre (P.) Ltd. v. Assistant Commissioner of Income-tax, Circle 12(2)

ORDER

K.K. Gupta, Accountant Member. - The assessee has filed this appeal agitating the action of the learned Commissioner of Income-tax (Appeals) in confirming the Arm's Length Price (in short 'ALP') as determined by the Assessing Officer in accordance with the Transfer Pricing Officer's Order.

2. The brief facts relating to the agitation can be stated as follows :—

2.1 During the previous year relevant to assessment year 2003-04, Philips Software Centre Private Limited (since merged with Philips Electronics India Limited) was engaged in providing software development services to its overseas affiliates. The assessee had various international transactions with its associated enterprises. The assessee conducted a transfer pricing study substantiating that its international transactions were at arm's length. The assessee filed its return of income for assessment year 2003-04 on 28-11-2003 disclosing a total taxable income of Rs. 2,97,86,420. Subsequently, assessment proceedings were initiated under section 143(3)/143(2) of the Income-tax Act, 1961 and the case of the assessee was also referred to the Additional Director of Income-tax (Transfer Pricing)-II (the TPO). The TPO passed an order under section 92CA of the Act, on 15-3-2006, proposing a transfer pricing adjustment of Rs. 2,21,80,792 after arriving at a mean margin of the comparable companies at the rate of 21.14 per cent. The assessment order under section 143(3) of the Act was passed on 31-3-2006 with the following primary adjustments.

(a) Reduction in the claim for deduction under section 10A of the Act to the extent of Rs. 1,31,26,943; and
(b) Transfer Pricing adjustment of Rs. 22,10,80,792 (as proposed by the TPO in his order dated 15-3-2006).

Being aggrieved by the assessment order, the assessee appealed before the CIT(A). The CIT(A), vide her order, dated 25-1-2008, has recomputed the mean margin of the comparable companies at 20.47 per cent and has revised the transfer pricing adjustment to Rs. 20,84,81,378. While passing her order and reaching her conclusions the CIT(A) has, inter alia, held that:—

(i) The TPO is well within his rights to conduct a fresh comparability analysis using contemporaneous data and the analysis conducted by the TPO was appropriate.
(ii) A mere difference in the rates of charging depreciation cannot be considered as a difference in accounting policies, requiring adjustments to arrive at an appropriate comparability. As such, making an adjustment using average rate of depreciation is not an appropriate approach.
(iii) Merely because a methodology for normalizing margins of companies earning super profits has not been laid down in the Act, it does not imply that the TPO cannot make an adjustment.
(iv) Risk adjustment of 5.25 per cent claimed by the company is without any basis and ad hoc, and cannot be allowed.
(v) Adjustment for working capital cannot be allowed merely on the basis that the same was allowed by the TPO in the succeeding assessment year.
(vi) The decision by the Hon'ble Delhi Tribunal in case of Mentor Graphics (Nokia) (P.) Ltd. v. Dy. CIT[2007] 109 ITD 101 is not applicable.

2.2 The assessee is in appeal before us in the following grounds :—

"The order passed by the ACIT under section 143(3) of the Income-tax Act, 1961 ('the Act') and the CIT(A) under section 250 of the Act, are bad in law and on facts.

The CIT(A) erred in determining the arithmetical mean of arm's length net profit margin of 20.47 per cent and a transfer pricing adjustment of Rs. 20,84,81,378.

The CIT(A) erred in law in not appreciating the meaning of the phrase 'having regard to' in sections 92 and 92CA(4) of the Act.

The CIT(A) erred in law confirming that the amendment made to section 92CA(4) vide the Finance Act, 2007 was to set right the lacuna in the said section.

The ACIT having accepted the international transactions of the appellant as being at arm's length in the assessment year 2002-03, the CIT(A) erred in law in confirming the different view adopted by the ACIT in the assessment year 2003-04.

The CIT(A) erred in law in holding that the Transfer Pricing Officer (in short 'TPO') could conduct a fresh comparability analysis at the time of assessment, using data which is not contemporaneous as per rule 10D(4) of the Income-tax Rules, 1962 (in short 'the Rules').

The TPO erred in not sharing the fresh comparability analysis conducted by him at the time of assessment and the CIT(A) erred in not commenting on the same in the impugned order.

The CIT(A) erred in facts in confirming the rejection of comparable companies selected by the appellant, without proving that the comparable companies were deficient or insufficient.

The CIT(A) erred in facts in confirming the comparability analysis conducted by the TPO, without having regard to the functions performed, assets employed and the risks assumed.

The CIT(A) erred in law and facts in disregarding the decision by the Hon'ble Delhi Tribunal in case of Mentor Graphics (Nokia) (P.) Ltd. v. Dy. CIT [2007] 109 ITD 101 .

The CIT(A) erred in law and facts in not granting the suitable adjustments to the net profit margins of the appellant vis-a-vis comparable companies, in terms of rule 10B of the Rules.

The CIT(A) erred in facts in normalizing super profit margins of comparable companies on an inappropriate basis.

The CIT(A) erred in law in not granting the benefit of +/-5 per cent variance as per proviso to section 92C(2) of the Act.

The CIT(A) erred in not commenting on the issue on exclusion of foreign currency expenditure and high speed link charges from the 'export turnover' while computing the deduction under section 10A of the Act.

The CIT(A) erred in (i) concluding that interest under section 234B of the Act is mandatory in nature; (ii) not holding that interest under section 234B of the Act should be computed on the basis of the returned income; and (c) not holding that interest under section 234B of the Act should be computed applying the rate of 1 per cent per month, being the rate prevailing on the date of assessment order.

The CIT(A) erred in confirming the levy of interest under section 234D of the Act."

3. The learned counsel initiating the arguments that have been covered in five to six sittings can be summarized as follows :—

3.1 'Philips Software Centre Private Limited' ('Philips Software' or 'the assessee' or 'the Company') was incorporated in India in September 1993 as a wholly-owned subsidiary of Royal Philips Electronics N.V., Netherlands. Philips Software subsequently merged with 'Philips India Limited', an existing 'Philips group company', with effect from 1-4-2004. The new merged entity is called 'Philips Electronics India Limited'. During the previous year relevant to assessment year 2003-04, Philips Software, had various international transactions which, inter alia, included rendering of software development services to overseas affiliates, and purchase of hardware/software.

3.2 The software developed by the assessee is only based on the instructions received from its associated enterprises. The assessee does not create/develop/sell software products and packages. The software deve-loped by the assessee is used by its associated enterprises for integrating the same with other software/hardware components developed by the latter. Therefore, the assessee's business is limited to providing contract software development services. The assessee conducted a Transfer Pricing Study ('TP Study') in order to comply with the Indian transfer pricing provisions. The assessee provides software development services to its associated enterprises in the areas of embedded and information system engineering, architecture design, programming and testing. The services are provided to the following six segments of the Philips group :—

(a) Philips Consumer Electronics;
(b) Philips Semi-conductors;
(c) Philips Research;
(d) Philips Centre for Industrial Technology;
(e) Philips Components; and
(f) Philips Intellectual Property and Standards.

Agreement and compensation :—

(a) The commercial terms and conditions of services between Philips Software and its associated enterprises are specified in 'Commercial Agreements'.
(b) Two sample commercial agreements are available at pages 152 and 155 of the paper book.
(c) During the relevant previous year, the assessee was remunerated on the following basis :—

(i) From 1-4-2002 to 31-12-2002 on a cost plus 5 per cent mark-up basis.
(ii) From 1-1-2003 to 31-3-2003 on a cost plus 10 per cent mark-up basis.

For the above purpose, 'cost' has been defined as all costs incurred by the assessee including/personnel cost, travel cost, infrastructure cost and depreciation.

(a) The assessee is a captive contract service provider, rendering software development services to its overseas affiliates. While rendering services to its overseas affiliates, the assessee does not bear significant business and operational risks.
(b) The assessee is insulated from the following business and operational risks :—

(i) Market risk.
(ii) Product risk.
(iii) Warranty risk.
(iv) Technology risk.
(v) Research & Development risk.
(vi) Intellectual property risk.
(vii) Credit and collection risk.

(c) The only two risks which the assessee bears, are manpower risk ( i.e., risk of losing its trained employees) and general business risk ( i.e., risk of not complying with the legal/statutory provisions of law). The effect of these risks is also minimized to a large extent as the assessee is remunerated for all costs with a mark-up.
(d) Irrespective of the final outcome of the software development activities undertaken, the assessee continues to be remunerated by their associated enterprises on a cost plus basis. Accordingly, the assessee is specifically insulated from the risk of under-utilisation of capacity, risk of re-work, and the other risks mentioned above.

As is clear from the above, the assessee does not bear significant business and operational risks and is only a captive contract software development service provider rendering services to its associated enterprises. The profile of the assessee clearly distinguishes it from full-fledged entrepreneurial companies. The assessee selected the CPM as the most appropriate method, inter alia, on the basis that :—

(a) the services were rendered by the assessee to its associated enterprises under long-term contracts;
(b) the costs incurred by the assessee were exclusively incurred for rendering the services.
(c) CPM would be appropriate considering the overall functions performed, assets employed, risks assumed and contractual terms; and
(d) the selection of CPM as the most appropriate method is also supported by the Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations issued by the OECD ('Organisation for Economic Co-operation and Development') which supports the use of the CPM under the following scenarios :—

- where related parties have concluded joint facility agreements or long-term buy-and-supply arrangements; or
- where the controlled transaction is the provision of services.

Under the CPM, the arm's length transfer price is determined by identifying the costs incurred by the service provider and adding an appropriate gross profit margin to the same. The gross profit margin is usually established with reference to comparable uncontrolled transactions. Further, to corroborate the results of the GPM, the assessee also used the TNMM in the TP Study as a supplementary analysis.

3.3 Generally, the TNMM is used when direct comparable are not available, because TNMM is an indirect method. Under this method, the net profit margin realised by an enterprises is computed in relation to some key factor such as sales, costs, assets, etc., and the ratio so computed is compared to the net margins realised by unrelated enterprises from comparable transactions. As the margins earned by the assessee were higher than those of the comparable companies, the international transactions of the assessee were at arm's length.

3.4 Once having selected the CPM as the most appropriate method, the assessee conducted a search process on the 'Capitaline 2000' data base ('Capitaline').

The said database is compiled by Capital Market Publishers India Limited and is a comprehensive interactive database of around 7000 Indian Companies, covering all companies listed on major stock exchanges like BSE/NSE plus other big unlisted companies. Based on the broad economic activity of the assessee, the 'computer software industry' was selected, wherein a set of 402 companies was generated. On this set of 402 companies, the assessee applied various quantitative (i.e., system based automatic elimination) and qualitative (i.e., manual elimination) filters as below :—

3.5 As is clear from the above table, out of the initial set of 402 companies generated :—

(a) 331 companies were screened out by the data base upon application of the quantitative filters without any human intervention; and
(b) 62 companies were screened out upon application of the qualitative filters.

3.5A As is clear from the above discussion, the filters applied by the assessee included;

(a) Companies whose turnover is less than Rs. 5 crores and greater than Rs. 250 crores were excluded.
(b) Companies carrying out diversified activities were excluded ( i.e., companies which were functionally not comparable were excluded); and
(c) Companies rendering predominant services to its related parties were excluded. As indicated above, the assessee had screened out companies having a turnover less than Rs. 5 crores and more than Rs. 250 crores. Companies having turnover less than Rs. 5 crores were screened out as being too small or companies operating at a mascent/start-up stage. On the other hand, companies having turnover of more than Rs. 250 crores were screened out on the basis that such companies would enjoy economies of scale on the basis of their size of operations. After the application of the filters, a set of 9 comparable companies remained. The remaining companies were not 'selected' by the assessee, but were those which survived a process of elimination.

3.6 The search process was conducted by the assessee during October, 2003, wherein this database as updated on 30-9-2003 was used. Accordingly, the data used for computing the profit margins of the comparable companies was the latest date which was available as on 30-9-2003. The due date for filing the income-tax return for assessment year 2003-04 was 30-11-2003. Accordingly the data used by the assessee was clearly complied with the requirement of rule 10D(4) which requires the data to exist by the specified date. For applying the CPM, the assessee computed the 'gross profit margins' (i.e., gross profit as a percentage of costs) of the comparables. For applying the TNMM, the assessee computed the 'net profit margins' (i.e., operating profit as a percentage of operating costs). In line with the provisions of rule 10B(1)(e)( iii), the assessee, while computing the gross profit and net profit margins of the comparables, made an adjustment to the depreciation charge of the comparable companies, The adjustment was made as the assessee charges depreciation on its assets at a much higher rate as compared to the comparable companies which survived the search process. In other words, the comparable companies charged depreciation on their assets at rates which were much lower than the rates followed by the assessee. As such, for an appropriate and equitable comparison at the gross/net profit margin level, there was a need to make the necessary adjustment for the difference in depreciation policy.

The following tabulation provides a brief of the comparables selected in the TP Study :—

Based on the above comparability analysis, since the margins (both gross and net) of the assessee were higher than that of the comparable companies, it was concluded that the transactions for sale of software development services by Philips Software to its Overseas affiliates were at arm's length.

3.7 The following details were filed before the TPO during the course of the proceedings before him :—

The TPO issued a show-cause notice on 3-2-2006, the highlights of the show-cause notice are as below :—

(a) Depreciation adjustment : The TPO did not agree with the depreciation adjustment made by the assessee in the transfer pricing study.
(b) New search for comparables : Without rejecting the choice of data base, comparability analysis or the most appropriate method selected by the assessee, the TPO conducted a fresh search.
(c) Methodology of the TPO : The TPO's search was conducted on a different data base, viz., Prowess. The search process included comparable companies having their turnover in the range of Rs. 100 crores and Rs. 250 crores. This search process generated a set of 20 companies, out of which, the TPO himself rejected six companies as not comparable to the assessee. Finally the TPO arrived at a set of 14 companies and compared their average net profit margin on cost with that of the assessee.
(d) Most appropriate method : The TPO selected the TNMM over the CPM, without providing any reasoning for doing so.
(e) TPO's methodology not known : The TPO did not provide the assessee with any reasons for ( i) using the Prowess; database ( ii) disregarding the most appropriate method as selected by the assessee and ( iii) disregarding the comparability analysis as conducted by the assessee. Also the TPO did not share the outputs of the search process, e.g., the list of companies eliminated at each step and the reason thereof, with the assessee.
(f) Based on his analysis as above the TPO computed a net profit margin (on sales) of the 14 comparable companies (as selected by him) at 18.40 per cent. On the basis of the above result, the TPO concluded that the arm's length price was different from that computed by the assessee.
(g) Updation of the comparables identified in the TP Study : The TPO finally updated the margins of the comparables identified in the TP Study with the latest data available ( i.e., updated with the data available after the specified date). He arrived at the net mean margin of such comparable companies at 16.82 per cent (over cost), computed without making any adjustment for depreciation. On the above basis, the TPO proposed to recompute the arm's length price by using the average margin of 16.82 per cent.

The assessee made detailed submissions against the show-cause notice, vide letter dated 21-2-2006.

3.8 The TPO passed his order on 15-3-2006. As against the show-cause notice where the TPO had proposed to recompute the arm's length price by adopting the arithmetical mean of the margins of the comparable companies as 16.82 per cent, in the order, the TPO computed the arm's length price by using the mean margin of 21.14 per cent.

3.9 In the order, the TPO dropped seven comparables proposed by him in the show-cause notice, and he arrived at a final set of seven comparable companies (including one company which had also been selected in the assessee's TP Study), and computed a net profit margin on costs of 21.14 per cent for these companies (i.e., the TPO applied TNMM in the order). Since the said margin of 21.14 per cent was higher than the corresponding profit margin of the assessee of 5.70 per cent, the TPO recomputed the arm's length price for software development services at Rs. 1,73,56,31,438. On this basis, the TPO proposed an adjustment of Rs. 22,10,80,792. The eight comparables rejected by the TPO were screened out for the following reasons :—

The 14 companies identified vide the TPO's search and their final acceptance/rejection matrix is summarized as under :—

The TPO finally computed the arithmetical mean of the seven comparables in the transfer pricing order at 21.14 per cent and recomputed the arm's length price for software development services at Rs. 1,73,56,31,438. He also then proceeded to propose an adjustment of Rs. 22,10,80,792, though the law requires the TPO to only compute the arm's length price. The Assessing Officer needs to compute the total income of the assessee having 'regard' to the arm's length price as determined by the TPO.

3.10 The TPO grossly erred in several areas of law and facts. Some of the relevant issues are summarized below :—

(a) The TPO selected the TNMM as the most appropriate method without providing specific reasons for rejecting the CPM, which was used by the assessee in the TP Study.
(b) The assessee had used Capitaline database for conducing a comparability analysis. The TPO used another database ( i.e., Prowess) without giving any reasons.
(c) The assessee had followed a methodical search process in the TP Study, and arrived at a set of comparable companies which were functionally comparable to the assessee. However, without highlighting any deficiency or insufficiency in the comparables, functionally or otherwise, selected in the TP Study, or the search process followed for arriving at those comparables, the TPO has rejected the comparables. In the above context it would be relevant to note that even in the show-cause notice issued, the TPO had proposed to recompute the arm's length price on the basis of the comparables in the TP Study. However, in the order, the TPO has used a set of comparable companies selected by him, the search process for which has not been shared with the assessee.
(d) Rule 10D(4) clearly specifies that the information relation to the international transactions, including the comparability analysis, has to be kept and maintained latest by the specified date, i.e., the date of filing of the income-tax return (in the instant case 30-11-2003). However, the TPO conducted a fresh comparability analysis during January and February 2006, using the date which was not available after the specified date.
(e) In the search process conducted afresh by the TPO, he has applied a turnover filter of Rs. 100 crores to Rs. 250 crores without providing sufficient reasons for applying such filters.
Though the TPO has not changed the upper end of the filter as applied by the assessee he has not explained why he considered Rs. 100 crores as an appropriate lower end of the filter. The lower end of the filter applied by him is less than the turnover of the assessee by approximately Rs. 50 crores, whereas the upper end of the filter as applied by him exceeds the turnover of the assessee by approximately Rs. 100 crores.
The TPO has indicated that the turnover filter is a very important criteria for selecting or rejecting comparables, showing complete disregard to the functional aspects of a comparable company. The TPO also appears to have not examined the asset base of the comparables as finally selected by him.
(f) The TPO has normalized the margins of the super profit making companies. The concept of 'normalization' is nowhere covered/included in the Indian Transfer Pricing provisions.
(g) The proviso to section 92C(2) provides that while arriving at the arm's length price, at the option of the assessee, a price which may vary from the arithmetical mean by 5 per cent can be considered as the arm's length price. Having computed the arithmetic mean of the arm's length price at Rs. 1,73,56,31,438, the TPO has not allowed the assessee the option of reducing it by 5 per cent as is required by law.
(h) Having accepted the value of the international transactions of the assessee to be at arm's length in the preceding assessment year, the TPO has taken a different view for assessment year 2003-04, completely ignoring that there was neither any change in the facts of the case nor in the nature of operations or terms of business of the assessee in assessment year 2003-04.

Further, during the assessment for assessment year 2002-03 the TPO had conduced the assessment by using the CUP method, whereas during the assessment for assessment year 2003-04, the TPO took a completely different view and selected the TNMM as the most appropriate method. In this regard, it is pertinent to note that the assessee has selected the CPM, supplemented by the TNMM in its TP Study for both assessment year 2002-03 and assessment year 2003-04.

3.11 The CIT(A) has observed that before the Assessing Officer the assessee had merely reiterated the submissions made before the TPO and no additional evidence was placed on record, and, accordingly, the Assessing Officer had no alternative but to complete the assessment based on the TPO's order. The CIT(A) further stated that the 'lacuna' in the Act has been 'set right' vide the Finance Act, 2007, by replacing the words 'having regard to' with 'in conformity with' in section 92CA(4) of the Act.

4. The learned Departmental Representative made submissions summarized as follows :—

4.1 The taxpayer has quoted two of the CBDT's circulars explaining the rationale for introducing the TP Provisions. The taxpayer has highlighted the part which says that "the basic intention underlying the new TP regulations is to prevent shifting out of profits by manipulating prices charged or paid in international transactions, thereby eroding the country's tax base". The taxpayer has emphasized the above per to develop an argument that if a particular company is enjoying tax-exemptions in India, there would be no motive on the part of such a company to shift its profit out of India and hence TP provisions need not be applied to the International Transactions undertaken by such a company. It was also argued by the taxpayer before the ITAT that in the case of the taxpayer the average rate of tax is much lower in India than the tax rate applicable to the AE in Netherlands. The issue has been dealt with by the CIT(A) in her order at para 7.1 of her order. The ITAT Spl. Bench has also decided this issue in the favour of the revenue in the case of Aztec Software & Technology Services Ltd. v. Asstt. CIT [2007] 294 ITR (AT) 32/ 107 ITD 141 (Bang.) (SB) . It is admitted that the TP provisions were introduced primarily to check shifting out of profits from India. But the incidence of profit shifting has to be determined as per the procedure laid down in the Act. As per the TP provisions if an international transaction is not at arm's length, an adjustment is required. Thus shifting of profits outside India has to be determined with reference to the ALP of the transactions. There is no need for the tax authorities to establish the motive or incidence of shifting of profits independent of ALP determination. The Act or the Rules framed in this regard do not mention the comparative tax rates applicable to the taxpayer and the AE respectively as a relevant point of reference before invoking the TP provisions or while determining the ALP. Besides, the taxpayer has taken a very simplistic view of the whole issue of tax avoidance. The taxpayer company avails tax benefits under section 10A. The exemption is not perpetual. It will lapse shortly. Thus keeping a long time perspective, the taxpayer would always arrange its transactions with the AE in such a way that in future also it does not have to pay taxes in India. Second it is worth noting that even section 10A companies are not exempt from other taxes such as dividend distribution tax. The taxpayer by not declaring proper profits in India is indirectly reducing its liability to dividend tax as well. In this connection, it would be of relevance to reproduce a portion of para 12 of the OECD Guidelines :—

"It is important to bear in mind that the need to make adjustments to approximate arm's length dealings arises irrespective of any contractual obligation undertaken by the parties to pay a particular price or of any intention of the parties to minimize tax. Thus, a tax adjustment under the arm's length principle would not affect the underlying contractual obligations for non-tax purposes between the associated enterprises and may be appropriate even when there is no intent to minimize or avoid tax. The consideration of transfer pricing should not be confused with the consideration of problems of tax fraud or tax avoidance, even though transfer pricing policies may be used for such purposes."

To sum up, there does not necessarily have to be any pre-conditionality of an intention to minimize or avoid tax in order to make an adjustment to the ALP and even though transfer pricing policies may be used for purposes of tax fraud or tax avoidance, there are other factors that need to be taken into consideration in dealing with transfer pricing.

4.2 With regard to the taxpayer's argument that when its income is exempt under section 10A, over-pricing or under-pricing of an international transaction would not affect the computation of its income, the Hon'ble ITAT held in the case of Aztec Software & Technology Services Ltd. (supra) that this argument is without force in view of the specific provisions contained in the first proviso to section 92C(4) wherein it has been clearly stated that no deduction under section 10A/10AA or section 10B or under Chapter VI-A shall be allowed in respect of the amount of income by which the total income of the assessee is enhanced after computation of income under this section. The Hon'ble ITAT's conclusion (para 21, page 33) is reproduced below :—

"It is abundantly clear that the Legislature while introducing the enactment did comprehend a situation requiring investigation and addition on account of computation of arm's length price in cases of the assessee entitled to benefit under section 10A/10AA or section 10B of the Act in the light of specific provision, it is difficult to contend that arm's length prices cannot be determined under section 92C or 92CA where assessee is entitled to benefit of above sections."

It was also argued by the taxpayer in the present case before the ITAT that the above observations in the Aztec Software & Technology Services Ltd.'s case (supra) hold good only up to the stage of reference to the TPO. The TPO should independently establish tax avoidance motive before initiating proceedings. This argument is correct neither in law nor on facts. The Tribunal's above mentioned observations are very clear that the TP provisions are equally applicable to tax exempt companies as to any other company. The findings given by the ITAT are applicable not only to reference under section 92CA but also to entire TP proceedings including determination of ALP by the TPO.

4.3 The learned counsel has argued that as per rule 10B(2), the assets employed is not an independent consideration in comparability analysis but only an important factor. The taxpayer is playing with words instead of putting any substantial argument. Rule 10B(2)(a) is very clear that the functions performed by the comparable companies should be compared taking into account the assets employed. The import of the rule is that even if the articles produced or the services rendered by two companies are similar in nature, the two may not be comparable depending upon the fact that the assets utilized by them are not similar. A fully automated manufacturing unit cannot be compared with a unit manually operated though both may be producing the same article. The nature of functions performed by any business enterprise actually depends on the assets employed. Functions performed by a handloom textile unit are different from those performed by a power loom textile unit, though ultimately both are manufacturing textile and operate in the same business segment. Thus the nature and quality of assets employed is very basic in determining functional comparability. In the case of Ranbaxy Laboratories Ltd. v. Addl. CIT [2008] 299 ITR (AT) 175 (Delhi), the Delhi Bench of the ITAT held that:—

"The analysis of comparison should consider total assets employed and assets used to earn profit." (p. 202).

4.4 The learned counsel's argued that there is no reference to 'value' of assets in the rule 10B(2)(a) again shows taxpayer's tendency to raise an argument for arguments sake. The comparability analysis under the TP provisions is data based. It is not based on physical inspection of the units to be compared. Apparently when rule prescribes that assets employed in any business are to be considered, one has to be go by the assets shown in the balance sheet. All the assets, whether tangible or intangible, are given a certain value in the Balance Sheet. This value may be cost or depreciated value, etc. The comparison will be based on such value. There is no other way the assets employed in business can be compared.

4.5 The issue of use of contemporaneous data under rule 10B(4) and existence of the data by the specified date under rule 10D(4) is discussed below separately. The taxpayer has only reproduced the provisions of the Act. However the issue of the role of the TPO versus the Assessing Officer is as below.

4.6 The learned counsel has argued that the captive service providers - the offshore software development centres of the multinational companies operate in a risk insulated environment. This is not entirely correct; though it is true that the risks undertaken by a captive service provider are different from those undertaken by an independent entrepreneur. The captive service providers have to bear the risks relating to dependence on a single customer. The concept of dealing with a single customer and the inherent risks involved have been discussed by M/s. Infosys in its audit report for the year 2004-05 as under (page 95) :—

"Client concentration.—We rely on repeat business based on the strength of our client relationships and a major portion of our revenues come from existing key clients. As the size of a client increases, it limits our pricing flexibility, strengthens the clients' negotiation capability, and reduces the ability to govern the relationship for mutual advantage. Also, the business growth of these large clients, their own profitability and changes in IT strategy have the potential to adversely impact our revenues and profitability and increase credit risk. However, large clients and high repeat business lead to predictable revenue growth and lower marketing costs. Therefore, to strike a balance, we have chosen to limit the revenue from any single client to a maximum of 10 per cent of total revenue."

4.7 Even in the stock market, shares of the software companies which are heavily dependent upon a very limited number of clients get lesser premium than the shares of the companies having multiple client base. Dependence upon a particular client or a particular geography alone is always considered a greater risk. Thus, doing business exclusively with a single client or a single group of clients carries different type of risk and therefore exclusivity clause in any contract is often accompanied by a balancing premium. The taxpayer is also entitled to a premium on account of single customer risk on the same lines as an independent entrepreneur would claim premium for bearing market risk.

4.8 There are factors at work that actually place a captive service provider in a much higher risk zone. For example, a captive service provider is completely dependent on its associated enterprise and any downswing impact in the business of the associated enterprise could have a severely damaging impact on the captive service provider. Such is not the case with an entrepreneur who has multiple clients and this diversification actually mitigates risk since a downswing in the business of one client results in a lesser impact on its business as compared to a captive service provider.

4.9 The claim that the captive service providers would be compensated all costs regardless of any work being clone or not is only being irresponsible. Every captive service provider normally has a contract with its AE to render the required services. The cost plus remuneration is assured only on rendering of the services to the satisfaction of the AE. No AE enters into a contract stipulating that the captive service provider would be compensated even if no services are rendered.

4.10 In many cases the captive service providers cannot seek business from any 3rd party without AE's prior written consent. This type of non-compete clauses puts heavy burden on an entrepreneur and normally warrant an extra compensation. No independent entrepreneur would agree to such a restrictive clause without being adequately remunerated.

4.11 The taxpayer's contract with the AE does not guarantee fixed volume of business for any fixed period. As per the Agreement dated 9-1-2003 between the taxpayer and Philips Industrial Activities Belgium, the validity of the agreement is only for a period of 3 months. The clause 4 of the agreement reads as under :—

"The commercial agreement covers the period from 1-1-2003 to 31-3-2003." (Copy of agreement is filed on pages 152 -155 of assessee's paper book).

4.12 As per the taxpayer the captive service providers work on a Cost Plus model of pricing which insulated them from up and down of the market rates. The claim is true only technically. In reality the AE is exposed to the market fluctuations and the terms of contract between the AE and the taxpayer are determined by the AE according to the market conditions. Further the volume of business for the captive service provider depends on the AE's business. If the AE's growth slows down, the flow of work to the AE also slows down. Thus indirectly the captive service provider is also affected by the market condition.

4.13 Besides, the mark up allowed is not uniform in the case of all the captive service providers. In some cases the mark up is allowed almost at par with the market rates whereas in some cases the mark up allowed by the AE ensures a profit rate which is much below the market price. Thus whether the mark up allowed is fair and equitable has to be seen from case to case.

4.14 There are many captive software; service providers who work in exactly similar business environment as the taxpayer. These companies also do not bear market risk etc., as claimed by the taxpayer. It is seen from the records that these companies are showing much better profit margins that the taxpayer. Following are few examples of such companies who disclosed substantially higher margins than the taxpayer in the financial year 2004-05 :—

1. CISCO Systems India - 50 per cent.
2. BEA Systems Customer Supp. & Research Centre - 30 per cent.
3. Core Objects India - 20 per cent.
4. ESI Software India - 20 per cent.
5. LG Software India - 28 per cent.
6. Logix Microsystems - 40 per cent.
7. Wipro Ltd. - 28 per cent.
8. Tektronics Engg. Ltd. - 40 per cent.
9. Mphasis BFL Ltd. - 27 per cent.
10. Subex Systems Ltd. - 32 per cent.
11. Sanyo LSI Tech. Ltd. - 21 per cent.
12. Silver Software - 35 per cent.
13. Manhattan Associates Ltd. - 29 per cent.
14. Snecma Aerospace Ltd. - 35 per cent.
15. Think 3 Design India - 40 per cent.
16. Torry Haris Business Soln. - 31 per cent.
17. Tavant Tech. Ltd. - 41 per cent.
18. MeCreade Software Ltd. - 92 per cent.
19. Multitech Software Systems - 25 per cent.
20. Spike Infotech - 30 per cent.
21. Magnasoft Cons. - 23 per cent.
22. S.G. Software Asia - 41 per cent.
23. Realsoft Ltd. - 31 per cent.
24. Relq. Software - 23 per cent.
25. Medicom Consultance Services - 83 per cent.

4.15 Similar cases are available for the financial year 2002-03 as well. This clearly shows that even a captive service provider can earn margins at par with the independently operated companies. The perceived risk premium is not really a hindrance in earning profits, particularly in the software sector. The software industry is not a capital intensive industry and does not involve a long gestation period etc.

4.16 The taxpayer has argued that the business risks are borne by the parent company and the Indian subsidiary is only a captive service provider with minimum risk. However, many studies have been conducted on MNCs setting their shops across the globe which have revealed that by multi-nationalising their operations, these MNCs are in fact decreasing their own risk. If the business risks pertaining to the captive service provider subsidiaries. Moreover, there should be increase in risk level for the MNC parent. However, many studies have been conducted to see how the degree of multinationality produce additional benefits in terms of excess returns or reduced risk and it is found that the degree of multinationality did not have a significant influence on the risk and return performance of the MNCs. (Degree of Multinationality and Financial Performance: A Study of U.S.-Based Multinational Corporations by Khursheed Omer, David Durr and Philip Siegel).

4.17 The taxpayer has also argued that the independent enterprises earn higher margins because they bear more risk. Thus according to the taxpayer there is a risk return trade off- higher the risk higher the margins. But this hypothesis does not appear to be true. It is already shown above that many of the captive service providers are earning margins better than the industry average. On the other hand there are many independent enterprises which have very low margins in spite of the fact that they bear all the risks associated with an independent entrepreneur, M/s. iGate Solutions, Lanco Global Solutions, Melstar Information Technologies, R.S. Software Systems etc., are some examples. Thus it cannot be said that higher risks always result into higher returns.

4.18 Independent enterprises have to incur expenditure on advertisement/marketing/selling etc., which is not required in the case of captive service providers. This adversely affects the comparable's margins. Therefore risk premium if any enjoyed by the comparable companies does not translate into absolute gains in the margins. The inclusion of expenditure on marketing, advertising etc., for the computation of the net margins, even out the gains by way of risk premium. This is also true that the risk borne by an independent entrepreneur results as frequently into losses as in the profits. Therefore the profit margin of an independent company cannot be considered as entirely dependent on the risk borne by that company.

4.19 The taxpayer had also selected independent enterprises like Visual Soft, Soft Cell, Lanco Global, Orient etc. as comparables in the TP study. The risk profile of the comparable companies selected by the taxpayer is the same as that of the comparables selected by the TPO. In the TP study the taxpayer neither claimed nor computed any risk adjustment. If the taxpayer's comparables can be accepted than the TPO's comparables should also be accepted because the risk profile of both the comparables set is the same.

4.20 The taxpayer has mentioned various risks such as market risk, product risk, credit risk, etc. However the taxpayer has not explained whether each of 9 comparables selected by him, bore all the risks enumerated in his submissions. As the comparables selected by the taxpayer had different risk profiles, the taxpayer should have explained which comparable bore what type of risks and how did the same affect its margins. The following table shows the expenditure incurred on selling and advertisement (marketing), R&D, etc., by the comparables selected by the taxpayer. The table also shows the bad debts written off or provided for in each case. It can be seen that there is no direct connection between the expenditure incurred on selling and marketing and the profit margin earned. Companies like Visual Soft and Infotech earned better margins than RS Software and Prithivi though the latter had spent more on marketing. Companies like Softsol, Lanco, etc., have not incurred any such expenditure still their margin is much higher. Thus it is clear that the profit margin of a company is not dependent on the marketing expenditure, etc.

4.21 The Act has put the primary onus on the taxpayer to determine the ALP and show that its international transactions are at arm's length. If any adjustments are claimed, it is for the taxpayer to quantify such adjustments. The taxpayer cannot use the claim of adjustments as a device to frustrate the TP proceedings. If, according to the taxpayer no quantification of the claimed adjustments is possible, the claim should be ignored because no adjustment: can be made without actual quantification. In the present case the taxpayer has not furnished any computation of the risk adjustments claimed. Hence the claim should be rejected. If the taxpayer furnishes at this stage any computation of the adjustments based on well-accepted principles, including the working capital adjustment, or if the Tribunal holds that some adjustment has to be allowed, the matter may kindly be restored to the file of the Assessing Officer/TPO for exact quantification of the risk adjustments including the working capital adjustment.

4.22 In the context of risk adjustments, it may also be kept in mind that the Act presumes that different comparables from the independent enterprises may have varying profit margins depending upon the Functions, Assets and Risk (FAR) profile of each comparable. The comparables selected by the taxpayer had margins ranging from -18.9 per cent (loss) on cost to 30.36 per cent profit on cost. To take care of such differences, the Act provides for adoption of arithmetical mean of varying profit margins of different comparables [proviso to section 92C(2)]. Thus the concept of arithmetical mean takes care of differences arising in the margins of the comparables on account of various factors including risk profile. Thus there are enough safeguards in the Act which takes care of the difference in profit margins of the comparable companies due to risk profile etc. The discussion on the risk analysis is summarized as under :—

The taxpayer is totally dependent on the AE for business. Thus the taxpayer takes the risks associated with heavy dependence on a single customer. In common business parlance it is known as 'single customer risk'.
The cost plus agreement with the AE does not guarantee sufficient volume of business for any definite period. The agreement can be terminated by any party at any time after giving a statutory period notice. Thus the taxpayers not free from the risk of losing business entirely or losing volume of business.
The AE is exposed to the market risk and any fluctuation in the; business conditions of the AE affect the contractual terms between the AE and the taxpayer.
There are many captive service providers earning much better margins than independent risk bearings enterprises and vice versa. Thus there is no direct correlation between the margins earned and risks taken.
Many business studies have shown that the MNCs are actually distributing their risks by opening captive offshore centres.
The independent entrepreneur has to incur expenditure on marketing, R&D etc., the expenditure is debited to the profit and loss account. Thus if undertaking the market risk etc., helps in earning any extra margin, the benefit is more than set off by the corresponding expenditure. The same applies to credit risk undertaken and the provision for bad debts.
The risk profile of the comparables selected by the taxpayer and those selected by the TPO is similar. The taxpayer cannot therefore ask for rejection of the TPO's comparables on this account.
The taxpayer did not furnish any computation of the risk adjustment either in the TP study or during the course of TP proceedings. The primary onus is on the taxpayer to compute the risk adjustments because he is claiming the same and also because the Act puts the responsibility of determination of the ALP primarily on the taxpayer.
Different comparables can have different risk profiles and different profit margins. The Act provides for adopting arithmetical mean of the different prices. This provision neutralizes the effect of difference in the risk profile and profit margins of various comparables.
Finally the taxpayer's argument that its low margins are linked to the low risk bearing nature of the enterprises is belied by the fact that the margins were changed during the year from 5 per cent to 10 per cent on cost. As per the agreement dated 1-1-2002 the taxpayer was allowed cost +5 per cent mark up, for similar service the taxpayer was allowed cost +10 per cent as per the agreement dated 9-1-2003. No special circumstances have been revealed by the taxpayer which warranted sudden increase in the margin from 5 per cent to 10 per cent despite the nature of the services to be rendered and the surrounding circumstances (FAR) remained the same as earlier. This is a very clear proof that the taxpayer had not fixed the price of international transactions at arm's length, but the same were fixed to suit the group's convenience or arbitrarily.

4.23 The taxpayer's claim of being free from various risks is not acceptable. As is discussed above the taxpayer is exposed to single customer risk. The taxpayer has just listed out all types of risk related terms which can be thought of. But the insinuation that the comparables selected by the taxpayer or the TPO bear all of these risks is not correct. All the selected comparables are contract service providers and hence there is no question of their having any produce risk or Research and Development risk or IP risk etc., basically there are two major differences in the risk profile of the comparables and the taxpayer.

4.24 The taxpayer does not have to indulge into marketing of its services. It can depend on the AE to provide it with business. This issue has already been discussed above that as against the market risk the taxpayer is exposed to single customer risk. The present taxpayer spends substantial amounts on advertisement and publicity (Rs. 62,77,152 in assessment year 2003-04 and Rs. 71,63,413 in assessment year 2002-03). This is inexplicable if the taxpayer is free from the market risk.

4.25 As regards the remuneration on cost plus basis it is pointed out that the taxpayer's agreement with the AE guarantees cost plus returns only with lot of strings attached. The agreement between the taxpayer and the AE as available on page 156 of the A's paper book shows that the maximum remuneration fixed was 57,000 Euro inclusive of 5 per cent margins for the work to be done from January to May 2002. Similarly the agreement between the taxpayer and the AE as available on page 153 of the A's paper book shows that the maximum remuneration fixed was 25,000 Euro inclusive of 5 per cent margins for the work to be done from January to March 2003. This cannot be considered as a free licence to the taxpayer to earn 5 per cent mark up on cost. As the upper limit of the expenditure to be reimbursed is fixed, the taxpayer cannot afford to underutilize its resources. The AE maintains a quality control and therefore normally there is no question of rejection of the services rendered. In any case there is nothing in the agreement between the AE and the Taxpayer that the AE would go or reimbursing the cost with 5 per cent mark up to the taxpayer even if the taxpayer fails to carry out the work to the AE's satisfaction.

4.26 The other difference is regarding credit risk. Normally speaking the taxpayer does not have a risk of bad debts or late recoveries because it is dealing only with the AE. This risk is taken care of by allowing working capital adjustment. The bad debts incurred by the comparables are any how debited to their P&L account and thus the profit margin gets automatically adjusted.

4.27 The taxpayer claims to have selected Cost Plus method primarily and TNMM as a supplementary. This issue is discussed below separately :

"The search process applied by the taxpayer was faulty because it chose a very arbitrary range of turnover for selection of the comparables. The filters applied by the taxpayer are discussed below in detail. The issue of use of contemporaneous data, application of CPM and the adjustment sought on account of depreciation are discussed below. No comments except that the comparables selected by the taxpayer were not suitable. No comments except that the conclusion drawn on the basis of a faulty TP analysis cannot be accepted."

4.28 The taxpayer's claim is wrong that the TPO carried out a fresh search without first rejecting the taxpayer's TP study. In the paras 1 and 2 of the show-cause letter dated 3-2-2006, (page Nos. 231-235 of the assessee's paper book) the TPO had clarified that the taxpayer had not applied CPM or TNMM in a proper manner. The adjustment asked on account of depreciation claim was not in accordance with the TP provisions. The TPO also clarified that the comparables selected by the taxpayer were not suitable comparables barring one comparable. As the TPO did not accept the Taxpayer's analysis, he had no option but to search for new and suitable comparables. The TPO invoked the provisions of section 92CA(3) in this regard which is perfectly justified.

4.29 In the show-cause letter the TPO had also clarified that looking to the availability of the data it was not possible to have a comparability analysis at Gross Profit level. Hence the TPO adopted PBIT as the PLI. It is factually incorrect to say that the TPO did not give his reasons for rejecting comparability at gross profit level which, as required under the Cost Plus method.

4.30 The TPO did not have to explain use of Prowess data base in the same way as the taxpayer was not required to explain use of Capitaline data base. The Act does not specify use of any particular data base. The TPO was required in the interest of natural justice to disclose the data used by him. This was done. The taxpayer cannot have any grievance. In any case the taxpayer is not arguing that the Prowess data base is not reliable or is not available in public. The TPO had very much disclosed the search process adopted by him. It is very clear from show-cause letter that the TPO as a first step had considered the companies, engaged in software development and having turnover between 100 cr. to 250 cr. The turnover limit was fixed considering the taxpayer's turnover. The taxpayer himself had applied a turnover criterion but with a difference that almost all the comparables selected by the taxpayer were many times smaller in size than the taxpayer. Out of the 20 companies so selected, the TPO rejected 6 companies for the rejections mentioned in the show-cause letter. The balance 14 companies were offered to the taxpayer for his comments. There cannot be a more transparent and objective search process. The turnover of the comparables selected by the TPO and the taxpayer are shown in the table below :—

It is apparent that the comparable selected by the TPO are turnover-wise closer to the taxpayer than the taxpayer's comparables.

4.31 The taxpayer's contention that the TPO had carried out a fresh de novo search of the data base is factually incorrect. In its reply dated 21-2-2006 (page 241 of the taxpayer's paper book) the taxpayer stated that all the comparables selected by the TPO had been examined by the taxpayer and eliminated from the comparables set in the TP study for various reasons. The TPO found that the reasons as cited by the taxpayer were not proper and the companies should have been selected as comparable. Thus the issue to be decided is suitability of the comparables selected by the TPO and whether rejection of the same comparables by the taxpayer in its TP study was justified. The TPO had not selected any 'new' company which had not earlier been considered by the taxpayer. Therefore in facts and circumstances of the present case the question as to whether the TPO was right in carrying out a fresh search of the data base is of only academic importance. It may also be pointed out here that the search process is not an end in itself. The suitable comparables can be selected in various ways viz., by search of the database, through company websites, publicly available annual reports of the companies, information available in industry surveys etc. The Act does not prescribe any one particular way of collecting the data or selecting the comparables. Ultimately every selected comparable should stand the test of FAR analysis. The taxpayer has to be provided with an opportunity to give his comments on the new comparables proposed by the TPO along with sufficient supporting data. So far as this requirement of natural justice is fulfilled, the search process or the source of the data used by the TPO is immaterial. In the instant case, the taxpayer has been provided with all the material gathered regarding the comparables selected/rejected. The taxpayer's arguments have been taken into consi-deration and each comparable has been finally selected or rejected after due FAR analysis.

4.32 In the show-cause letter the TPO had made different propositions to the taxpayer. The TPO had mentioned in para 1.4 that the taxpayer's own comparables had an average (PBIDT) profit margin at the rate of 35.20 per cent as against the taxpayer's PBDIT margin of 18 per cent. Hence adjustment was required. The next proposition of the TPO was that the 14 new comparables selected by the TPO had PBIT margin at the rate of 18 per cent as against the taxpayer's PBIT margin of 6.61 per cent. Hence adjustment was required. The third proposition of the TPO was that the comparables selected by the taxpayer had shown PBIT margin at the rate of 16.83 per cent on cost as against the taxpayer's PBIT margin of 6.61 per cent. Hence adjustment was required. The taxpayer was requested to comment on the entire three scenarios. After considering the taxpayer's objections the TPO selected the final set of the comparables and worked out the ALP accordingly. The taxpayer cannot complain of lack of opportunity or violations of principles of natural justice.

4.33 The taxpayer has objected that the TPO having determined the ALP proceeded to propose the adjustment under section 92CA. It is seen from the order that the TPO did not 'propose' any adjustment. He simply computed the amount of the adjustment consequent to the determination of the ALP. How the taxpayer is prejudiced by this is not clear. The computation of the adjustment is a simple mathematical function. The TPO simply mentioned the difference between the ALP determined by him and the price charged by the taxpayer as the adjustment amount. This was done to double check the figures mentioned in the final order.

4.34 On the issues arising out of the TPO's order, the ld. DR submitted that he had resorted to selection of TNMM without rejecting the method adopted by the taxpayer. The taxpayer had selected cost plus method in the TP study. The taxpayer had also used TNMM as an alternative. The TPO adopted only TNMM. Both the methods are profit based. The cost plus method compares the companies at gross profit level whereas TNMM compares at the net profit level.

4.35 For applying CPM, the taxpayer has selected 9 comparables and compared their gross margins after adjustment towards depreciation. The Gross Profit Margin was computed by reducing the cost of services from the consideration received (Annexure 6 of TP Report). The following expenses of the taxpayer were considered as direct and indirect costs of services :—

Employee cost
Travelling and conveyance
Communication expenses
Purchase of software and
Depreciation

The amount spent on Rent (9.48 crores), Power, water and fuel (2.35 crores), insurance (0.52 crores), repairs (5.09 crores), lease and hire charges (Rs. 1.8 crores) and professional charges (Rs. 3.54 crores) were not considered as part of the services. This obviously resulted into a higher gross profit margin for the taxpayer.

4.36 The TPO rejected the CPM for the following reasons mentioned in the show-cause notice dated 3-2-2006 (page 234 of the Paper Book submitted by the taxpayer) as under :—

"it may be mentioned here that for the purpose of computing gross margin, accurate data is required in respect of direct and indirect costs and other costs in respect of all the comparables chosen. Since such data is not available operating profit to operating cost/operating profit to operating sales is adopted considering all costs (except interest) i.e., PBIT as PLI."

The stand taken by the TPO is correct in law because as per rule 10B(1)(c), the CPM is to be applied as under :—

"10B. (1) For the purposes of sub-section (2) of section 92C, the arm's length price in relation to an international transaction shall be determined by any of the following methods, being the most appropriate method, in the following manner, namely :—

(c)cost plus method, by which,—

(i) the direct and indirect costs of production incurred by the enterprise in respect of property transferred or services provided to an associated enterprise, are determined;
(ii) the amount of a normal gross profit mark-up to such costs (computed according to the same accounting norms) arising from the transfer or provision of the same or similar property or services by the enterprise, or by an unrelated enterprise, in a comparable uncontrolled transaction, or a number of such transactions, is determined;
(iii) the normal gross profit mark-up referred to in sub-clause ( ii) is adjusted to take into account the functional and other differences, if any, between the international transaction and the comparable uncontrolled transactions, or between the enterprises entering into such transactions, which could materially affect such profit mark-up in the open market;
(iv) the costs referred to in sub-clause ( i) are increased by the adjusted profit mark-up arrived at under sub-clause ( iii);
(v) the sum so arrived at is taken to be an arm's length price in relation to the supply of the property or provision of services by the enterprise;"

As mentioned in the above rule, all the direct and indirect costs incurred in providing services by the taxpayer should form part of the cost. However, the taxpayer has ignored expenditure such as professional charges, rent, repairs and other costs which are incurred in providing services by the taxpayer. Further, all the costs except interest is towards provision of services by the taxpayer is also clear from the following extract of the agreement entered between the taxpayer and its AE, Philips Industrial Activities, N.V. Belgium (dated 1-1-2002).

"The Billing pattern every month will be as follows :—

(a )Invoice value will be arrived by adding 5 per cent margin to the total cost incurred for a particular month....Total cost shall mean all costs incurred by Philips Software Centre Private Limited in this regard, and shall specifically include the personnel costs, travel costs, infrastructure costs (including depreciation on capital equipments)."

From the above it is clear that all infrastructure related costs like insurance, repairs and maintenance, lease and hire charges towards equipment and rent form part of costs in rendering services by the taxpayer. In para 4.2-1(b) of the written submissions before the Tribunal the taxpayer has categorically stated that, "The costs incurred by the appellant were exclusively incurred for rendering the services,". Still the before mentioned expenses have been excluded by the taxpayer while arriving at the cost. So, the TPO correctly rejected the CPM method and applied TNMM method as all the costs except interest go into providing services by the taxpayer. Even, the OECD guidelines say as under in regarding the Cost Plus Method :—

"Cost plus method.—A transfer pricing method using the costs incurred by the supplier of property (or services) in a controlled transaction. An appropriate cost plus mark up is added to this cost, to make an appropriate profit in light of the functions performed (taking into account assets used and risks assumed) and the market conditions. What is arrived at after adding the cost plus mark up to the above costs may be regarded as an arm's length price of the original controlled transaction.

2.36 The cost plus method presents some difficulties. In proper application, particularly in the determination of costs. Although it is true that an enterprise must cover its costs over a period of time to remain in business, those costs may not be the determinant of the appropriate profit in a specific case for any one year. While in many cases companies are driven by competition to scale down prices by reference to the cost of creating the relevant goods or providing the relevant service, there are other circumstances where there is no discernible link between the level of costs incurred and a market price (e.g., where a valuable discovery has been made and the owner has incurred only small research costs in making it).

2.39 Another important aspect of comparability is accounting consistency. Where the accounting practices differ in the controlled transaction and the uncontrolled transaction, appropriate adjustments should be made to the data used to ensure that the same type of costs are used in each case to ensure consistency. The gross profit mark ups must be measured consistently between the associated enterprise and the independent enterprise. In addition, there may be differences across enterprises in the treatment of costs that affect gross profit mark ups that would need to be accounted for in order to achieve reliable comparability. In some cases it may be necessary to take into account certain operating expenses in order to achieve consistency and comparability, in these circumstances the cost plus method starts to approach a net rather than gross margin. To the extent that the analysis takes into account operating expenses, the reliability of the analysis may be adversely affected, for the masons set forth in paragraphs 3.29-3.32. Thus, the safeguards described in paragraphs 3.34 - 3.40 may be relevant in assessing the reliability of such analysis.

2.40 While precise accounting standards and terms may vary, in general the costs and expenses of an enterprise are understood to be divisible into three broad categories. First, there are direct costs of producing a product or service, such as the cost of raw materials. Second, there are indirect costs of production, which although closely related to the production process may be common to several products or services (e.g., the costs of a repair department that services equipment used to produce different products). Finally there are the operating expenses of the enterprise as a whole, such as supervisory, general and administrative expenses.

2.41 The distinction between gross and net margin analyses may be understood in the following terms, in general, the cost plus method will use margins computed after direct and indirect costs of production, while a net margin method will use margins computed after operating expenses of the enterprise as well. It must be recognised that because of the variations in practice among countries, it is difficult to draw any precise lines between the three categories described above. Thus, for example, an application of the cost plus method may in a particular case include the consideration of some expenses that might be considered operating expenses, as discussed in paragraph 2.39. Nevertheless, the problems in delineating with mathematical precision the boundaries of the three categories described above do not alter the basic practical distinction between the gross and net margin approaches."

From the above, the CPM method cannot be applied to the taxpayer's case for the following reasons as mentioned in the above OECD guidelines :—

(a) The CPM method presents some practical difficulties in identifying the costs incurred for the provision of services like whether an indirect cost is towards rendering services or it is an enterprise level expense.
(b) There is no discernible link between the level of costs incurred and a market price in software services as the services are usually compensated on a man hourly basis which may not vary much across the industry for same or similar type of service. So, CPM is not the appropriate method.
(c) While applying CPM, the taxpayer should have considered all direct and indirect costs incurred in respect of the services rendered by it. But, it is apparent from details given above that the taxpayer itself did not include all the direct and indirect costs while computing gross mark up e.g., rent, insurance, repairs, lease/hire charges on equipment were not considered. Professional charges were also excluded without giving any reasons. In the case of comparable companies, the details of indirect costs incurred in rendering services were simply not available. Thus it cannot be said that the gross profit worked out in the case of the comparables is after taking into account the same items of expenditure as in the case of the taxpayer.
(d) As per the agreement, the taxpayer's cost of providing services includes all costs that have been incurred by the taxpayer except interest expenses. All these expenses are reimbursed on cost plus basis. The taxpayer is forgetting that it is also getting 5 or 10 per cent on other costs which have been excluded by the taxpayer in CPM like rent, repairs, insurance, lease and hire charges etc. So, the gross margin shown by the taxpayer distorts the true picture of its financial for which TNMM method is the most appropriate method as it captures all expenses except interest, coinciding the cost base of the taxpayer for reimbursement (as per the agreement and invoices raised by the taxpayer).
(e) In the taxpayer's case all the costs were incurred towards rendering of services. Therefore the gross profit is the same in the taxpayer's case as the net profit. Thus the taxpayer's PLI remains the same whether CPM is applied or the TNMM is applied. The difference between the two methods in the taxpayer's case is thus of only academic interest.

Hence, the TPO is correct in rejecting the CPM method and applying the TNMM method as the taxpayer itself had applied the same as an alternative analysis. Looking at the deficiencies of data for CPM, the TPO had taken TNMM as the Most Appropriate Method (MAM). The reasons for rejection of cost plus were conveyed to the taxpayer vide the show cause letter mentioned above.

4.37 It is worthwhile to reproduce the observations of the Hon'ble Supreme Court in the case of Dit (International Taxation), Mumbai v. Morgan Stanley & Co. Inc..) [2007] 292 ITR 416 , where the Court held as under :—

"The taxpayer is required to compute the arm's length price for a transaction(s) using one of the five methods stipulated in the Income-tax Rules. Rule 10C(1) of the Income-tax Rules defines the most appropriate method as the method which is most suited to the facts and circumstances of each particular international transaction. As per rule 10C(2) the most appropriate method has to be selected having regard to a number of factors which are enumerated therein."

4.38 The Apex Court has clearly given primacy to rule 10C(2) for purposes of selecting the most appropriate method, which is wider in scope than rule 10B(2) and has been sought to be relied on by the taxpayer in favour of its arguments against the application of a turnover criteria by the TPO. The Hon'ble Court further observed as under :—

". . . the methods quoted above namely CUP, RPM, CPM, PSM &TNMM are mentioned in section 92C read with rule 10B. The most appropriate method has to be applied for computation of the arm's length price. It will depend on facts and circumstances of each particular international transaction. . ."

The Supreme Court has also held TNMM to be the most appropriate method in the case of Service PE - "in our view apart from the orders passed by the TPO/Assessing Officer the said method (TNMM) is the appropriate method in the case of service PE as the TNMM apportions the total operating profit arising from the transactions on the basis of sales. Quite clearly, the use of the TNMM method by the TPO has the support of the highest court of the land.

4.39 Selection of a different database by the TPO as is already mentioned above, the Act does not specify use of any particular database. It is not the taxpayer's case that the Prowess data base used by the TPO is less reliable than the Capitaline database used by the taxpayer. Both the data bases are publicly available and contain details about Indian companies, the taxpayer has also not given any reasons for selection and use of Capitaline in place of Prowess data base in the TP study. In any case, no new comparables were considered by the TPO. The comparables selected by the TPO had already been considered and rejected by the taxpayer in its TP study, (refer para 5.2 above).

4.40 The taxpayer had selected comparables which were different from the taxpayer in size and asset base. The taxpayer cannot argue that turnover and size of a company is not an important factor in determining comparability. The taxpayer himself had selected companies having turnover between Rs. 5 crore to Rs. 250 crore 116 companies were excluded by the taxpayer applying the turnover criterion. Companies like Infosys, Satyam, L&T Infotech were excluded by the taxpayer because of their size and turnover. If turnover is not an important factor than the taxpayer would not have excluded these companies. Even now the taxpayer is not agreeing that these companies should be included in the comparable set.

4.41 Thus there is no argument on the principle that turnover should be retained as a criterion. The only argument is about whether the turnover range applied by the TPO is better than the turnover range selected by the taxpayer. The taxpayer's turnover was Rs. 151 crore in the relevant year. The taxpayer had applied a turnover range of Rs. 5 crore to Rs. 250 crore however in practice the taxpayer preferred the smaller companies. Out of the 09 comparables selected by the taxpayer only 01 comparable had turnover exceeding Rs. 100 crore. All the other comparables had turnover less than Rs. 100 crore 02 comparables had turnover between Rs. 100 crore and 50 crore. The balance 06 comparables had turnover less than Rs. 50 crore the smallest comparable had turnover of only Rs. 7.41 crore. The TPO found that the taxpayer's selection of the comparables was skewed.

4.42 The TPO therefore applied a minimum turnover limit of Rs. 100 crore and maximum turnover limit of Rs. 250 crores as a criterion. The taxpayer had taken upper side turnover limit of 160 per cent of the taxpayer's turnover of Rs. 151.75 crores. The TPO therefore applied the same limit on the upper and lower both side i.e., companies having less than 60 per cent of the taxpayer's turnover, rounded off to Rs. 100 crore, were also excluded. The taxpayer has also argued (Page 31 of letter dated 12-3-2008 filed before ITAT) that the TPO accepted the higher side limit of Rs. 250 crores i.e., 100 crores more than that of the taxpayer's, whereas on the lower side, the limit is only Rs. 100 crores i.e., Rs. 50 crores less than that of the tax payer's. This interpretation of the taxpayer is misleading. The argument is unsustainable because in percentage terms, the limit applied by the TPO on higher side as well as lower side is more or less identical 60 per cent less or 60 per cent more than the turnover of the taxpayer. Further, the upper limit is the same as applied by the taxpayer itself.

4.43 In this context reference may be made to the order of the ITAT, Pune Bench 'A' in E-gain Communication (P.) Ltd. v. ITO [2008] 23 SOT 385 . In this case the taxpayer had a turnover of Rs. 10.25 crore the taxpayer had selected comparables having turnover between 8 crore to 18 crore, the TPO selected comparables having turnover between 8 crore to 364 crore and the CIT(A) applied a turnover range of Rs. 5 crore to 25 crore for selection of comparables. The ITAT upheld the CIT(A)'s action. Applying the ratio of the ITAT decision it can be said that the turnover range applied by the TPO in the present case gave better comparables.

4.44 The taxpayer's argument that the TPO did not examine the asset base of the comparables is also wrong and self-contradictory. The most important asset in software industry is the employees, the human capital. As the employee cost of the comparables companies was in the same range as that of the taxpayer, the asset base of the comparable companies was also comparable to that of the taxpayer. Moreover, the turnover in general is directly proportional to the number of employees as the billing in the software industry in general is based on man hourly basis. Therefore, turnover gives a close approximation to the number of employees as the data regarding the number of employees is not available in respect of all comparable companies selected by the TPO in public domain. Thus the turnover limits applied by the TPO threw up closer comparables and hence, should be accepted as correct.

4.45 In the show-cause notice the TPO had pointed out that even the comparables selected by the taxpayer were showing higher margins than the taxpayer. This does not mean that the TPO had accepted the taxpayer's comparables. In fact the same show cause letter mentioned the reasons why the taxpayer's comparables deserved to be rejected.

4.46 The filters applied by the taxpayer were not proper filters as can be seen below :

(a) The taxpayer excluded companies having significant manufacturing activity without clarifying the meaning of significant.
(b) Similarly a turnover limit of Rs. 1 crore to 250 crore was applied in an arbitrary manner. Out of the 09 comparables selected by the taxpayer except for M/s. Visual Soft all other comparables did not have even half of the taxpayer's turnover. Thus the selected companies were not comparable in size and turnover to the taxpayer.
(c) The taxpayer excluded 36 companies as they were carrying out diversified activities. But no where is it clarified as to what constitutes diversified activities. Companies such as Geometric Software, Infotech Enterprises etc. were eliminated though these companies are mainly into software development.
(d) The taxpayer excluded 08 companies as they were rendering predominant services to their associated enterprises. However the taxpayer has not clarified the meaning of predominant. Melstar Inf. selected by the taxpayer had 50.12 per cent related party transaction in financial year 2002-03 and 17 per cent RP transactions in financial year 2001-02. Similarly M/s. Softsol had almost 100 per cent transactions with the related parties in the financial year 2002-03 as well as financial year 2001-02. The taxpayer should not have selected these companies as comparable as per its own filter.
(e) The taxpayer excluded companies incurring loss at the gross level because they do not represent the industry trend. However for the same reasons companies incurring loss at net level should also have been excluded but this was not done.

4.47 The comparables selected by the taxpayer had following defects :

(i) M/s. Melstar Inf. had significant related party transactions. In the financial year 2002-03 the company had sold services worth Rs. 4.60 crore to its associate concerns. Similarly the company's overseas subsidiaries had been paid service charges to the extent of Rs. 18.18 crore both these transactions put together constitute 50.12 per cent of company's total turnover. The company had about 17 per cent RP transactions in the financial year 2001-02.
(ii) M/s. Melstar Inf. had a turnover of 45.29 crore in the financial year 2002-03. The total cost incurred by the company was Rs. 43.30 crore which included software development expenses incurred outside India of Rs. 25.91 crore. Thus more than 50 per cent of total cost was incurred outside India. The company was not predominantly an offshore software development company is also clear from the fact that its total expenditure on employee cost was only Rs. 6.13 crore about 13 per cent of the total sales. Melstar is thus not a pure offshore development centre. More than 50 per cent of its turnover pertains to the job work got done abroad which is represented by the development charged paid in foreign currency. The taxpayer is rendering only offshore services and thus Melstar cannot be accepted as a suitable comparable.
(iii) M/s. Softsol is one of the comparable selected by the taxpayer. The company is a subsidiary of US based Softsol Research Inc. Softsol India's entire sales were to the US Company in the financial year 2001-02 and also in the financial year 2002-03, this company cannot be accepted as a comparable.
(iv) M/s. Orient Inf. is also not an offshore service provider like the taxpayer. Orient is basically an ONSITE service provider which is apparent from the following details of the employee cost incurred by the company. The living allowance paid to the consultant amount to 82 per cent of the total cost (Rs. 25.87 Crore). The onsite nature of the company is also clear from the balance sheet of the company. An onsite service provider company cannot be compared with the taxpayer who is an offshore service provider. The onsite work is done at the clients place i.e., out side India. The employees have to be paid in line with salary structure of that country. This practically wipes out the margins resulting from the low salary structure prevailing in India. Thus an onsite service provider company works at relative disadvantageous position. M/s. Orient should not have been selected as a comparable.
(v) M/s. S.M.R. universal had total revenue of Rs. 4.65 crore in financial year 2002-03 and employee cost of only Rs. 0.18 crore i.e., less than 5 per cent of the revenue. No software company can work without employees. The employee strength is reflected in the employee cost. The taxpayer had 46 per cent of its revenue spent on employee cost. Apparently SMR did not have enough employees to indulge in software development. The major expenditure in the case of SMR is other operating expenses, though full details of such expenses are not available in the data base. The company deserves to be excluded from the comparable set.

These issues were not examined and discussed in detail by the TPO in his order as most of the taxpayer's comparables failed the turnover filter itself.

Contemporaneous data - The issue has been discussed in detail in CIT(A)'s order pages 26 to 31.

The issue of using relevant financial year data and using multiple year data was considered by the Hon'ble Special Bench of ITAT in the case of Aztec Software & Technology Services Ltd. (supra ) upholding the revenue's objection against use of multiple year data by the taxpayer, the Hon'ble Tribunal held as under (page 140, para 223) :

"It cannot be said that this base data was relevant and sufficient to come to any findings about the arm's length price. The objection of the revenue based on sub-rule (4) of rule 10B for not confining to data for the financial year 2001-02 is also justified as it is not shown that data for other years had any influence on the determination of ALP of the transactions involved."

The Special Bench decision was followed by the ITAT, Delhi Bench in the case of Mentor Graphics (Nokia) (P.) Ltd. (supra). The relevant part of the order is reproduced below :

"The ld. CIT(A) was of the view that taxpayer was not justified in taking into account data for the earlier two preceding years. In his view, only the data for the current year should have been taken into account. He has given detailed reasons in different paras. We are not recording all reasons/details as during the course of hearing before us, ld. Representative of the taxpayer agreed that arm's length pricing has to be determined by taking only the data of the current year. He fairly conceded that the Special Bench in the case of Aztec International has also taken the same view and the said view has to be followed by this regular Bench."

This view is reaffirmed in para 35.2 of the order as under :

"35.2 Further there is contradiction in the approach of TPO; whereas he rightly insisted in the light of rule 10B(4) that only data for the financial year was relevant but for his own use, he has taken into account data for financial year 2003-04 i.e., relating to two years later."

It can be seen that the TP analysis involves three different stages.

(i) First stage is when the taxpayer enters into the agreement with the AE and the actual international transaction takes place. At this time the taxpayer determines the transfer price as per the prevailing business conditions and economic environment and the groups TP policy. At this stage the taxpayer does not have the contemporaneous data available.
(ii) The second stage is when the taxpayer undertakes the TP analysis as required by the Act and files the 3CEB report. This is a post facto exercise. The taxpayer has to use the contemporaneous data as available to him at the time. The provisions of rule 10D(4) are applicable at this stage. But at this stage the taxpayer is not permitted to use earlier year's data unless the conditions mentioned in the proviso to rule 10B(4) are satisfied.
(iii) The third stage in the TP analysis is when the TPO takes up the TP audit. The TPO is also bound to use the contemporaneous data. However there is no restriction in the Act or the Rules that the TPO should confine himself to those data which was available to the taxpayer at the time of 3CEB audit. The TPO is expected to use all the material and facts available to him at the time of TP audit.
(iv) To sum up, the TPO was correct in law and facts in using the current year's data for comparability analysis.

4.48 The issue of the turnover filter has already been discussed above. Apparently the turnover range adopted by the TPO gave closer comparables than the one selected by the taxpayer. It is wrong on the part of the taxpayer to argue that the TPO displayed total disregard to functional comparability. Except for the turnover criterion the TPO had used the same filters as the taxpayer. Aberrations by the TPO in applying the related party filter have been corrected by the CIT(A). The comparables selected by the TPO were not lacking in functional comparability is clear from the detailed discussion in the CIT(A)'s order regarding FAR analysis of each of the TPO's comparable.

4.49 The learned Departmental Representative further submitted that normalisation of the margins was undertaken by the TPO in favour of the taxpayer. The original margins of the companies may be restored at the insistence of the taxpayer. The TPO's order does not get vitiated because the TPO adopted an approach beneficial to the taxpayer, the normalization done by the TPO is actually a kind of adjustment provided for in rule 10B. TP is still a new subject in our tax administrator. It is quite possible that different persons dealing with the provisions may resort to certain practices in the interest of natural justice, though these practices may not find a place in the letter of the law.

4.50 The taxpayer has raised the issue that the TPO erred in not applying the proviso to section 92C(2) of the Act in allowing variation of 5 per cent from the arithmetic mean. The CIT(A) has discussed this issue at para 11 page 38 of her order.

4.51 The CIT(A) has discussed about TPO's treatment in earlier/current year at para 10 (Page 37) of her order. As held by CIT(A) in her order, res judicata is not applicable to income-tax proceedings including transfer pricing proceedings. This legal position has been confirmed in various judicial decisions. One of the recent judgments on the issue is the order of the ITAT, Lucknow 8 Bench in the case of Jaiprakash Investment (P.) Ltd. [It Appeal No. 787/Luck. of 2005, order dated 25-1-2008]. The determination of ALP is a fact based exercise. The selection of the most appropriate method and suitable comparables depend on the availability of facts. The availability of facts may differ from year to year and depending on the availability of facts different methods may also have to be chosen in different years. The facts in the case of selected comparable companies may also change. A company may not have related party transactions in one year, but may have them in the next year. The companies may change their business or may get acquired by different companies etc. All these changed facts would have a bearing on selection of comparables. Therefore it is quite likely that some of the comparables chosen in the earlier year may not qualify the test of uncontrolled transaction in the succeeding year. Thus in TP proceedings it is not possible to apply res judicata.

4.52 Besides, the original order for the assessment year 2002-03 has been set aside by the CIT as the same was found erroneous. The taxpayer cannot argue that an erroneous order has also to be followed. As is discussed by the CIT(A) in her order and also in the order passed by the CIT under section 263 for the assessment year 2002-03, the CUP is not proper method to be applied in the taxpayer's case. This view is also upheld in the case of Aztec Software & Technology Services Ltd. (supra) by Hon'ble ITAT, Bangalore wherein the TPO has used CUP method and adopted the average billing rate shown by NASSCOM. The matter is restored to the Assessing Officer by the Tribunal as under :—

"161. Having noted that hourly rate of billing would vary depending upon qualification and experience level of personnel, and having noted the different clauses of rule 10C(2) requiring consideration of specific characteristics of the property, functions performed, assets employed, risk assumed and other conditions of comparability, the Transfer Pricing Officer in the computation did not adopt or attach any importance to special characteristics or conditions of contract, and wrongly relied upon article 9 of the OECD Model Convention which had no application in the matter.

169. Rule 10A(d) also provides that a transaction includes a number of closely linked transactions. But here it is nobody's case that all controlled transactions carried for different customers were closely linked transactions and similar was the position of comparable uncontrolled transaction. In fact, by taking average, services of a totally different person were taken into account.

170. In the light of above discussion, we set aside the impugned order of the Commissioner of Income-tax (Appeals) and restore the matter to the file of the Assessing Officer who may again refer the question of determination of Arm's Length Price to the Transfer Pricing Officer. In our view, Transfer Pricing Officer should use his power and first call upon the taxpayer to furnish ALP and all material and information which he is obliged to maintain under rule 10D of the Income-tax Rules."

4.53 Thus it is clear that the TPO's order passed in assessment year 2002-03 was erroneous and unsustainable in law. The order has already been set aside by the competent authority. Basically the taxpayer has raised the issue as to whether the Assessing Officer is bound by the order of the TPO passed under section 92CA, the same is discussed below.

4.54 He argued on the difference between 'having regard to' and 'in conformity with' before and after amendment to section 92C(4). The CIT(A) has discussed this issue at para 7 (page 16) and para 12 (page 40) of her order. In this connection, the observations made by the Hon'ble ITAT Bangalore Special Bench in the case of Aztec Software & Technology Services Ltd. (supra) in para 51, page 67 are of relevance and are therefore reproduced below :—

"We are, therefore, of the opinion that TPO's order under sub-section (3) of section 92C cannot be treated as final and binding on the Assessing Officer. Having said so, we do not mean to suggest that order of TPO is of no value or consequence. The words "having regard to" convey definite meaning and are strong enough to enjoin upon the Assessing Officer to pass order and adopt transfer pricing as determined by the TPO unless there are very good grounds to modify or alter the transfer pricing ordered by the TPO. Only after recording reasons, as above, the Assessing Officer can take the transfer price other than one determined by the TPO. There may be cases in which the assessee, after arms' length price is determined by the TPO and before his order is made the basis of assessment, may get such authentic material to show that transfer pricing determined by the TPO is not correct or should not he blindly adopted without modifying it. However, in the cases before us, the TPO adopted the orders of the TPO and computed assessments accordingly. We see no illegality in the procedure followed by the TPO and therefore do not approve of the contrary observations of the ld. CIT (Appeals) in the impugned order."

The Tribunal had also considered meaning of the words 'having regard to' and the amendment to section 92C(4) in its order at Paras 52, 53 and 54 as under :—

"52. As far as decision of Supreme Court in the case of Rajesh Kumar v. Dy. CIT 287 ITR 91 = (2006-TIOL-140-SC-IT) is concerned on which strong reliance was placed their Lordship, as already noted, have observed that the expression 'having regard to' in this context assumes some significance. An opinion must be formed strictly in terms of the factors enumerated therein.

The expression indicates that in exercising the power regard must be had also to the factors enumerated therein together with all factors relevant for exercise of that power.

Their Lordship further quoted the following extracts from two decisions of Supreme Court in the above report :

(i) India Cement Ltd. v. Union of India [1990] 4 SCC 356.
'The meaning of the expression 'having regard to' is well-settled. It indicates that in exercising the power, regard must be had also to the factors enumerated together with all factors relevant for exercise of that power.'
(ii) Delhi Farming and Construction (P.) Ltd. v. CIT [2003] 260 ITR 561,

'The words 'having regard to' used in the section do not restrict the consideration only to two matters indicated in the section as it is impossible to arrive at a conclusion as to the reasonableness by considering only the two matters mentioned isolated from other relevant factors. It is neither possible nor advisable to lay down any decisive tests for the guidance of the Income-tax Officer. The satisfaction depends upon the facts of each case. The only guidance is his capacity to put himself in the position of a prudent businessman or the directors of a company and his sympathetic and objective approach to the difficult problem that arises in each case.'

It is clear from above that decision of Supreme Court in the case of Rajesh Kumar (supra) itself indicate that words 'having regard to' suggest that assessment is to be made having regard to the report of the TPO which is required to be considered with other relevant material available on record.

There is nothing to suggest that TPO's report on transfer pricing is conclusive and debars Assessing Officer from looking at any other material.

The aforesaid conclusion is also in line with latest change made in section 92C by the Legislature through the Finance Act, 2007. Sub-section (4) of section 92CA has been substituted with the following sub-section with effect from 1-6-2007 :

'[(4) On receipt of the order under sub-section (3), the Assessing Officer shall proceed to compute the total income of the assessee under sub-section (4) of section 92C in conformity with the arm's length price as so determined by the Transfer Pricing Officer]'

53. Now words 'having regard to' have been replaced by words 'In conformity with'. So now Assessing Officer after introduction of sub-section (4) above is required to pass assessment order in conformity with the order of the Transfer Pricing Officer determining arm's length price: Now the order of the TPO has been expressly made binding on the Assessing Officer. From the above it is clear that there was a lacuna in the Act as appropriate language was not used earlier. This has been modified and with effect from 1-6-2007, the order of TPO is binding on the Assessing Officer who now has no choice but to pass an order in conformity with the order of the TPO. The words "having regard to" did not convey the same meaning. For all the aforesaid reasons, we hold that prior to substitution of sub-section (4) by a new section, the order of the TPO was not binding on the Assessing Officer.

54. . . . Ordinarily without strong reasons, the Assessing Officer should not challenge the transfer pricing determined by the TPO and bring to naught his efforts after receipt of the reference. Our observations that it was not binding on the Assessing Officer, prior to 1-6-2007 are meant to convey that the assessee even before the Assessing Officer could file evidence and show that arm's length price shown by the assessee is quite reasonable and should be accepted. This question is decided accordingly."

In the instant case the Assessing Officer after receiving the TPO's order allowed an opportunity of hearing to the taxpayer. The Assessing Officer also considered the arguments filed by the taxpayer. The taxpayer had not raised any fresh objections before the Assessing Officer. Naturally the Assessing Officer had no reason to disturb the ALP determined by the TPO. The Assessing Officer's action cannot be held to be in violation of the principles of the natural justice or against the provisions of the Act.

4.55 The taxpayer has referred to CBDT Circular 14/2001 and argued that ALP determined by the taxpayer should not be disturbed. The said circular actually does not put any such restrictions on the TPO. The circular in fact re-emphasize the legal position that if based on the information or material furnished by the taxpayer or collected by him, the TPO is of the opinion that ALP has not been determined correctly, he can compute the ALP as per the provisions of the Act and the manner prescribed in the Rules.

The relevant portion of Circular No. 14/2001 is reproduced below :

"Under the new provisions the primary onus is on the taxpayer to determine an arm's length price in accordance with the rules, and to substantiate the same with the prescribed documentation. Where such onus is discharged by the assessee and the data used for determining the arm's length price is reliable and correct, there can be no intervention by the Assessing Officer. This is made clear by sub-section (3) of section 92C which provides that the Assessing Officer may intervene only if he is, on the basis of material or information or document in his possession, of the opinion that the price charged in the international transaction has not been determined in accordance with sub-sections (1) and (2), or information and documents relating to the international transaction have not been kept and maintained by the assessee in accordance with the provisions contained in sub-section (1) of section 92C and the rules made thereunder; or the information or data used in computation of the arm's length price is not reliable or correct; or the assessee has failed to furnish, within the specified time, any information or document which he was required to furnish by a notice issued under sub-section (3) of section 92D. If any one of such circumstances exists, the Assessing Officer may reject the price adopted by the assessee and determine the arm's length price in accordance with the same rules. However, an opportunity has to be given to the assessee before determining such price. Thereafter, as provided in sub-section (4) of section 92C, the Assessing Officer may compute the total income on the basis of the arm's length price so determined by him."

The Board's Circular referred to by the taxpayer also states that the tax-payer's ALP can be accepted only if the data used is reliable and correct. In this case, it has been demonstrated by the TPO as well as by the CIT (A) that the data used by the payer was not correct. The circular did not grant an amnesty to the taxpayers.

4.56 The taxpayer had eliminated 08 companies from the comparable set on account of their having controlled transactions. The taxpayer had applied a filter of eliminating companies with predominant related party transactions. However in the TP study the taxpayer did not elaborate further on this filter. As mentioned above one of the comparables selected by the taxpayer Melstar Ltd. had RP transactions amounting to 50 per cent of the company's turnover and another of taxpayer's comparable Softsol had almost all its sales to the related party only. Thus there was no consistency in the taxpayer's approach on the issue of related party transactions. The TPO also applied a filter of substantial related party transactions. The following companies falling within the turnover range applied by the TPO were excluded on account of substantial related party transactions :

Mphasis BFL Ltd.

Mastek Ltd.

Pentasoft Technology.

The taxpayer argued before the CIT(A) that the following out of the comparables selected by the TPO should be excluded on account of related party transactions:

Flextronics Software

Hinduja TMT

Infotech Enterprises

Silverline Technologies

4.57 The CIT(A) has held that companies having more than 25 per cent of their total operating income as related party transactions should only be excluded. Companies having RP transactions below this limit can be retained as comparable. The CIT(A)'s order has been accepted by the department. The CIT(A) deleted only one company - Flextronics Software from the list of comparables.

4.58 The Act requires that the controlled transactions should be compared with uncontrolled transactions. In common parlance uncontrolled transactions means a transaction between two unrelated parties. The Act also permits aggregation of the transactions for the purpose of comparability if the transactions are closely linked. In the TP study done by the taxpayer all the transactions of the taxpayer have been clubbed to arrive at the profit margins. Similarly in the case of comparable companies also all the transactions pertaining to software development services have been clubbed together. In other words as the tested party as well as the comparable companies are in the software development service the enterprise level profits have been compared. Looking to the nature of the business i.e., software development service this approach of the taxpayer appears to be correct and acceptable.

4.59 However, when comparing profits at enterprise level and clubbing together all the transactions a question arises about the treatment to be given to some of the transactions which the comparable company might have carried out with its associate concerns and other related parties. Practically speaking almost every company or business concern would have related party transactions such as salaries paid to Directors, sale purchase with associated concerns etc., though the quantum and effect of such transactions on the overall margins may vary from case to case. If such transactions with the related parties are significantly affecting the overall enterprise level profits of that company, then certainly that company would have to be excluded from the list of comparable. On the other hand if the related party transactions constitute a minor and insignificant part of the overall turnover, then the same may not affect the overall margins of the company and hence that company may still be retained as a comparable subject to functional similarity.

4.60 To overcome the difficulty of eliminating companies with significant controlled transactions but at the same time retaining an adequate number of comparables it was thought proper to exclude all those companies who had related party transactions in excess of 25 per cent of their operating revenue.

4.61 Though it may be ideal to select only those companies which have Nil related party transactions, it would not be a practical approach because almost all the companies have related party transactions to some extent. Nil criteria would result into elimination of all the companies or selection of very few and very small companies which may not give sufficient base for comparison.

4.62 The taxpayer has also impliedly accepted that companies having related party transactions above certain limit only should be excluded. The taxpayer had applied a filter of companies having 'predominant' related party transactions. Though the taxpayer did not specify any particular limit in per cent terms, it is apparent that the taxpayer is also against elimination of companies having insignificant or relatively few transactions with the related parties.

The ITAT Special Bench has held in the case of Aztec Software & Techno-logy Services Ltd. (supra) (para 223 page 140) that an adequate number of comparable companies should be selected for analysis. To quote from the order;

"The sample size of data taken by the taxpayer, to support the arm's length price, was too small to come to any general conclusion. The assessee had taken profitability figures of only ten US companies and even it was not established that activities of these ten companies were same and similar, i.e., to provide onsite and marketing services on cost plus markup basis, it cannot be said that this base data was relevant and sufficient to come to any findings about the arm's length price. The objection of the revenue based on sub-rule (4) of rule 10B for not confining to data for the financial year 2001-02 is also justified as it is not shown that data for other years had any influence on the determination of ALP of the transactions involved."

4.63 In the case of Aztec Software & Technology Services Ltd. (supra) providers the transactions are so interlined that they have to be aggregated to determine the profit margins. Thus the enterprise level profit margins have to be considered in the place of transaction level profit. The aggregated transactions at the enterprise level may include some transactions with the related parties as well. However if the related party transactions do not have material effect on the overall profit margins then still that company can be considered as a comparable. The Act does not provide directly as to what per cent of related party transactions can have material effect on the overall margins. But however some guidance can be taken from definition of the Associated Enterprise in section 92A(2)(a) where in it is prescribed that one enterprise holding 26 per cent shares in the other enterprise can be considered as an associate enterprise. Similarly in the provisions of section 40A(2)(b) the persons having substantial interest is described as a person carrying not less than 20 per cent of voting power in that company. Thus it is seen that 20 per cent or 26 per cent interest is considered as substantial interest. As the provisions of section 92A(2)(a) are from the transfer pricing chapter itself, a limit of 25 per cent may be considered more reasonable to be applied as the threshold limit for the related party transactions. If the limit is reduced further it would only result in eliminating more and more companies, on the other hand if the limit is relaxed then companies with predominantly related party transactions would get included which would not represent uncontrolled transactions. Therefore on balance 25 per cent appears a proper threshold limit for related party transactions. The companies having more than 25 per cent related party transactions should only be rejected as comparables.

The OECD guidelines regarding adoption of enterprise level profits in TNMM and treatment of partial controlled transactions are reproduced below:

"3.42 An analysis under the transactional net margin method should consider only the profits of the associated enterprise that are attributable to particular controlled transactions. Therefore, it would be inappropriate to apply the transactional net margin method on a company-wide basis if the company engages in a variety of different controlled transactions that cannot be appropriately compared on an aggregate basis with those of an independent enterprise. Similarly, when analysing the transactions between the independent enterprises to the extent they are needed, profits attributable to transactions that are not similar to the controlled transactions under examination should be excluded from the comparison. Finally, when profit margins of an independent enterprise are used, the profits attributable to the transactions of the independent enterprise must not be distorted by controlled transactions of that enterprise."

The filter of 25 per cent related party transactions have been adopted because transactions with related parties exceeding 25 per cent of total turnover may distort the overall profit margins of that company. While calculating the 25 per cent, the transactions on account of sales as well as purchases have been combined and percentage on the overall sales of the comparable company worked out.

The rule 10B(1)(e)( iii) prescribes that under TNMM

"the net profit margin referred to in sub-clause (ii) arising in comparable uncontrolled transactions is adjusted to take into account differences, if any, between the international transaction and the comparable uncontrolled transactions, or between the enterprises entering into such transactions, which could materially affect the amount of net profit margin in the open market."

And rule 10B(3)(i) prescribes that —

"none of the differences, if any, between the transactions being compared, or between the enterprises entering into such transactions are likely to materially affect the price or cost charged or paid in, or the profit arising from, such transactions in the open market;"

Related party transactions less than 25 per cent of the turnover cannot materially affect the overall profit margin of the company. Hence such companies can be selected as comparable under TNMM.

4.64 The TP provisions or the OECD guidelines do not prescribe profit margin as a comparability criterion. No comparable should be rejected or accepted on the basis of its profit margins. Comparability analysis should be based purely on similarity of the nature of services rendered or goods sold/purchased, functions, assets employed in the business, economic conditions etc. therefore the taxpayer's argument that companies with higher margins should have been excluded is not in accordance with the TP provisions. The TPO had resorted to normalization of the profits of some of the comparables only in order to take a liberal view in the whole matter. To some extent the normalization of profits compensated for various risk adjustments asked for by the taxpayer. The taxpayer should not have any grievance. There was a substantial difference between the employee cost to sales ratio in the case of the taxpayer and in the case of the two companies whose profits were normalized. The taxpayer has a sales to employee cost ratio of 45.1 per cent (employee cost/sales) as against the same in the case of Hinduja TMT @ 20 per cent and in the case of Infotech Enterprises @ 34.10 per cent. The process of normalization was therefore required. Moreover it is also mentioned here that M/s. Infotech had only 31.78 per cent margin on cost which is only slightly higher than the normalized margin @ 28.61 per cent (the margin in the case of Visual Soft).

4.65 The CIT(A) has done a detailed FAR analysis of the comparables selected by the taxpayer. The taxpayer's objections regarding the following comparable companies were considered by the CIT(A) and she found that the companies are functionally similar to the taxpayer.

(i) Prithvi Information Solution Ltd.
(ii) Xansa India Limited.
(iii) Infotech Enterprises.
(iv) Hinduja TMT Limited.

The taxpayer's objection that the company's export turnover was less than 75 per cent of the total turnover was rejected by the TPO because in the relevant year the company's exports constituted more than 75 per cent of its turnover. The details are reproduced below from the Prowess data base:

Clearly, the main activity of the company is software, particularly export. It can be seen that the taxpayer's objection is factually incorrect.

It is also seen that the company was rendering services in 3 segments namely, Information Technology, IT enabled Services and Media and Telecommunication, its expertise in software services was in offering end-to-end computing solutions on various platforms :—

- IT Consulting
- Architecture and integration
- Customized Services Development
- Application re-engineering and Migration
- Application maintenance
- Training

The company apparently belongs to the software industry.

The taxpayer has taken an argument repeatedly that the comparables selected by the TPO were engaged in diversified activities. The taxpayer has referred to various verticals of software development in which these comparable companies are engaged. Here we may see that the companies; selected by the taxpayer were also engaged in diversified activities and different verticals of software. Hence verticals of software was never a criterion of comparability either in the taxpayer's selection or the TPO's selection. Thus the TPO's comparables cannot be found fault with on this account. The comparables selected by the taxpayer were engaged in diversified activities as mentioned below.

Lanco Global renders services in the following sectors :

SOLUTIONS

IT Governance, Compliance & Security Services
Enterprise Resource Planning
SAP
Oracle
eBusiness Applications
Customer Relationship Management
Business Intelligence
Dataware house
Implementations
Application Development
Upgrade & Migration
Application Management Services
Infrastructure Support Services
Testing & QA
Business Process Outsourcing

M/s. Melstar has the following to say about its functions:

Melstar Technology

An Application Management Company

"Everyday Melstar is helping its customers across the globe in building and managing their business applications by virtue of its domain knowledge in Banking, Insurance and Information Technology as well as its Technical Expertise in IBM, Microsoft and Sun Technologies."

Services

MelSupport MelSupport, our application support service, offers services that can be undertaken onsite, off-site, offshore, or any combination thereof. Ongoing application maintenance and support and Integration, enhancement and support of third party applications.
MelSolutions MelSolutions, our Flexible Engagement Model provides assistance in : Applications Development, Applications Management, Product based Solutions.
MelSoft

MelSoft works as an extension of your own application development, support and maintenance centre.

MelTest
MelConsult

Orient information technology renders the following services;

"With a broad mix of functional and industry experience, Orient's technical resources can work on various levels such as software developers, project managers, database specialists, networking, support executives and others. They can be deployed on projects of all leading technologies on Client-server and Internet such as Web development initiatives, Complex e-businesses, Telecom and Wireless Solutions, Enterprise wide applications, CRM Solutions, Database Solutions, Software Migration and Maintenance Services.

Orient designs, develops and deploys customized software solutions and applications. Our list of offerings include -

New and Turnkey development

Re-engineering, Customization, and Migration of legacy applications

1. Upgrades and Enhancements in functionality
2. Application Migration to new software languages
3. Application Porting to new technologies and platforms

Our practices focus on implementing solutions like SAP, Peoplesoft, Oracle and others. We provide entire gamut of consulting, development and implementation services on a wide range of solutions and packages. Combine this with our experience in understanding various industry verticals as well as XRM processes; and we have the right ingredients to ensure a definite ROI."

M/s. R.S. Software is in the following fields:

Application Development & Maintenance Testing & QA Managed Services Staff Augmentation

Thus it can be seen that the companies selected by the taxpayer were also engaged in diversified fields of software development. The same criterion was applied by the TPO.

Adjustment on account of Depreciation, the issue had been discussed by the CIT(A) at para 8.7 of her order.

4.66 The taxpayer asked for adjustment on account of depreciation on the ground that the taxpayer had charged depreciation at higher rate than the comparables. This is not acceptable for the following reasons :—

A. Under the TNMM this type of adjustments are not required. OECD guidelines describe the strength and weakness of TNMM as under:

"One strength of the transactional net margin method is that net margins (e.g., return on assets, operating income to sales, and possibly other measures of net profit) are less affected by transactional differences than is the case with price, as used in the CUP Method. The net margins also may be more tolerant to some functional differences between the controlled and uncontrolled transactions than gross profit margins. Differences in the functions performed between enterprises are often reflected in variations in operating expenses. Consequently, enterprises may have a wide range of gross profit margins but still earn broadly similar levels of net profits.

The taxpayer has also given the same reason while selecting TNMM as an alternative method at page 62 of the TP study (page 87 of the taxpayer's paper book).

B. The taxpayer has asked for an adjustment based on the average rate of depreciation. The average rate has been worked out by dividing the total depreciation claim by the total WDV of all the fixed assets. In the case of the taxpayer the opening WDV of the assets as on 1-4-2002 was Rs. 20,15,47,185 during the year assets worth Rs. 13,71,63,863 were added whereas assets worth Rs. 34,35,335 were deleted. Thus the total works out to Rs. 33,52,75,713. The total depreciation claimed by the taxpayer during the year is 14,26,11,707 which works out to 42.53 per cent of the total assets. While computing the adjustment, the taxpayer applied this average rate of depreciation @ 43 per cent on the opening net block of assets of all the comparable companies. This computation is incorrect and misleading because different class of assets (land, plant and machinery, buildings, office equipment etc.) carry different rate of depreciation. In the taxpayer's own case he has charged depreciation @ 12.5 per cent only on leasehold premises. The taxpayer has charged higher rate of depreciation on computer equipments and plant and machinery etc. The average rate of depreciation would be thus higher in the cases where plant and machinery and computers etc. constitute major part of the overall block of assets. Conversely the average rate would be lower if assets like building, leased premises, furniture and fixture etc. constitutes major portion of the block of assets. For example we can see that in the case of Soft Sol, one of the taxpayer's comparables, the average rate of depreciation (depreciation Claim/total value of assets) comes to only 15 per cent (1.06/7.02). However, further scrutiny reveals that this company has also claimed depreciation @ 50 per cent on software, @ 40 per cent on computers and so on. The average rate of depreciation is so low because (the major part of the block of assets is building worth Rs. 2.59 cr. on which the depreciation claim is only Rs. 12.91 lakh at the rate of 5 per cent. Apparently the so called average rate of depreciation is not necessarily always indicative of the fact that depreciation has been charged at lower rates. It can just indicate the nature of the assets on which depreciation has been charged. As the depreciation rates for different class of assets are different, one uniform rate cannot be applied to all the assets. The adjustment worked out by the taxpayer on the basis of average rate of depreciation is thus not acceptable.

C. The taxpayer has consistently applied depreciation at the rate of 43 per cent on a straight line method. Consequently, the closing WDV of assets in the case of the taxpayer is reduced to that extent, whereas in the case of other companies who are following lower depreciation rates, the decrease in the value of assets each year is also less to that extent and consequently the closing WDV of assets remain relatively higher. In other words, though the taxpayer had applied depreciation at higher rate, the base (opening WDV of assets) is much lower, whereas in the case of comparable companies depreciation though applied at a lower rate, is applied on a higher base. Thus the effect of higher depreciation rate is considerably reduced. If any adjustment/normalization of depreciation has to be worked out, it must be worked out minimum for the last three years, so that the base effect can be neutralized.

D. One way to tackle the problem of differences in depreciation rates is to take PBDIT (Profit Before Depreciation, Interest and Tax) as the PLI. The profit before depreciation will rule out any effect on the margins on account of difference in the rate of depreciation. The TPO had done this exercise and found that the taxpayer was earning PBDIT at the rate of 18 per cent on cost whereas the average PBDIT of comparables selected by the taxpayer's comparables was at 23.98 per cent on cost which falls within +/-5 per cent limit of the taxpayer's margins. Thus as far taxpayer's comparables are concerned it does not make any difference whether comparison is done at the PBIT level or PBDIT level. Even if the TPO's comparables are taken the difference is not much whether PLI is fixed at PBIT level or PBDIT level. The difference between the taxpayer's margin and the arithmetical mean of the comparable companies' margins does not change much. If comparison is 'done at the PBDIT level, the difference works out to 28.62 per cent on cost and if the comparison is done at PBIT level the difference works out to 30 per cent.

Thus depreciation can be excluded in the software services sector as the human resources rather than the fixed assets drive its revenues. In this respect, the Draft Notes issued for public comment by the OECD on Transactional Profit Methods (TNMM in the Indian context) also supports the view that the PLI can be considered before depreciation as well in the industries which are not asset intensive like software services industry. The relevant portion is reproduced as under:

Depreciation and amortisation

136. Whether or not to include depreciation and amortisation costs depends on the circumstances of the case. Depreciation and amortisation charges are very susceptible to differences in accounting treatment and in particular in the deemed accounting life of the asset. This may materially distort the comparison between the tested party's results and those of third party comparables in asset-intensive industries and depending on the circumstances of the case a more reliable comparison might be achieved using a net profit margin ratio that does not include depreciation and amortisation.

137. It was for instance suggested that when companies have significant acquisitions that result in the revaluation of both tangible and intangible assets to 'market' value, Earnings Before Interest, Tax, Depreciation and Amortisation (EBITDA) may be the most reliable measure, and that it may be appropriate to exclude book intangible assets from the analysis, as such book intangible assets are typically the result of acquisitions, and do not include internally developed intangibles.

138. Depreciation and amortisation can pose difficult comparability issues, for instance if two parties own and/or use comparable intangibles, but one has acquired them and amortised them over time, while the other has developed them and expensed them upfront.

139. For this reason, some commentators suggest that depreciation and amortisation ought to be excluded from the determination of the net profit margin indicator. On the other hand, in asset-intensive industries where assets are key value-drivers, excluding depreciation and amortisation might not lead to a meaningful outcome, and depreciation and amortisation would not be excluded if it can be reasonably assumed that they do not create material comparability issues. Where uncertainties of that type are material, the third party comparable concerned might have to be rejected. Where no or insufficient satisfactory comparables are available to apply the considered net margin indicator, another net margin indicator that is less sensitive to depreciation and amortisation costs might need to be considered."

Rule 10B(1) provides for adjustments only if the alleged difference between the taxpayer and the comparables affect the margins. In the instant case the result of the comparability analysis remains the same whether we take PBIT or PBDIT as PLI. This shows that the issue of adjustment to the margins of the comparables on account of depreciation is not the core issue. As the claim of adjustment itself is disputable and it is not possible to work out exact adjustments, the comparable companies PBDIT margins may be compared with the taxpayer's PBDIT margins. The working is given in the TP order (Page No. 249 of the taxpayer's Paper Book).

The Depreciation as a percentage of cost of the comparables selected by the TPO, as confirmed by CIT(A), and also that of the taxpayer are given below:—

As can be seen from the above; the percentage of depreciation of the taxpayer as a percentage of total cost (before interest, tax and other non-operative expenses) does not vary much from average of the same ratio for the comparable companies. The PBDIT margins of the comparables selected by the TPO, as confirmed by CIT(A), are given below:—

Thus even if depreciation is ignored and PBDIT is taken as the PLI, even then the taxpayer's transactions are not at arm's length and a transfer pricing adjustment is justified.

4.67 As per rule 10B(1)(e)( i), the net profit margin earned by the tested party from an international transaction is computed in relation to the costs incurred or sales effected or assets employed or having regard to any other base. The provisions of rule 10B(1)(e)( i) also allow selection of any appropriate base for comparing the margins. Thus in the taxpayer's case PBDIT can be adopted as PLI, if the taxpayer still feels that the depreciation is making a significant effect on the margins.

4.68 The taxpayer has referred to the decision of Pune ITAT in the case of Egain Communications. In the said decision, the Tribunal had not actually approved the computation of the adjustment. The Tribunal merely held that such an adjustment may be granted in the facts of the case. The Tribunal however did not say that average rate of depreciation should be applied to the comparable companies. On the contrary, the Tribunal held that the taxpayer's (tested party) margin may be reworked out by applying lower depreciation rates as per the Companies Act. This may be applied in the present case provided the taxpayer can establish that the finally selected comparables as approved by the Hon'ble ITAT had applied lower rates as per the Companies Act. In the case of EGAIN COMMUNICATIONS, the actual computation of adjustment on account of depreciation rates was not argued before the Tribunal. The decision cannot in any way be interpreted to be supporting the adjustments computed by the taxpayer in the present case.

4.69 The taxpayer had claimed before the CIT(A) that the difference between the PLR (the rate at which the banks give loans to customers) and the bank rate (rate at which RBI pays to the banks) should be considered the risk premium. Thus analogy is not supported by any logic. It is the taxpayer's understanding that the difference between the bank rate and the PLR represents risk premium. This is not correct. The RBI has always explained that its functions are guided by the economic policy. The RBI decides the bank rate, CRR, repo Rate etc. depending upon the economic conditions. The commercial banks decide their PLR keeping in mind the cost of the funds, NFA etc. As a practice all major commercial banks usually have the same benchmark PLR. But it is a wrong conception of the taxpayer that the banks give loan to their customers at the PLR. PLR is more like a reference point. The actual interest rates depend upon the nature of the loan such as personal loans usually carry a higher rate of interest than home loans. The actual interest rate depends on need for the bank spread, competition in the market and the Govt.'s policy on sectoral loaning. The taxpayer's notion that the difference between the bank rate and the PLR represents risk premium is totally misconceived. The taxpayer's claim of risk adjustment based on such misconceived notions cannot be accepted.

4.70 It may also be mentioned here that banking industry is entirely different from the software industry to which the taxpayer belongs. The risks undertaken by a banking company are of different nature. There are separate laws to regulate functioning of the banking companies. The economic factors which affect the banking industry are different from the economic factors which affect the software sector. In such circumstances no parallel can be drawn between the alleged risk premium charged by the banks and the risk premium which should be available to a captive service provider in the software sector in India.

4.71 Working capital adjustment - The taxpayer did not compute the same during the TP proceedings. If the Hon'ble ITAT decides that the same has to be allowed the matter may be remanded to the Assessing Officer/TPO for actual computation.

4.72 The CIT(A) has discussed the applicability of Mentor Graphics (Nokia) (P.) Ltd.'s case (supra) decision in her order in detail. In brief,

(a) The ITAT upheld the use of contemporaneous data. The para 19 of the order dealing with contemporaneous data is reproduced below:
"The ld. CIT(A) was of the view that taxpayer was not justified in taking into account data for the earlier two preceding years. In his view, only the data for the current year should have been taken into account. He has given detailed reasons in different Paras. We are not recording all reasons/details as during the course of hearing before us, ld. Representative of the taxpayer agreed that Arm's length pricing has to be determined by taking only the data of the current year. He fairly conceded that the Special Bench in the case of Aztec International has also taken the same view and the said view has to be followed by this regular Bench"
This view is reaffirmed in para 35.2 of the order as under:
"35.2 Further there is contradiction in the approach of TPO; whereas he rightly insisted in the light of rule 10B(4) that only data for the financial year was relevant but for his own use, he has taken into account data for financial year 2003-04 i.e., relating to two years later."
In the instant case the TPO used the contemporaneous data whereas the taxpayer had used earlier year's data.
(b) In para 39.4 of its order the ITAT held that unless it is shown that the TP analysis carried out by the taxpayer is insufficient or has some deficiency the TPO could not reject the study and carry out a fresh search. The relevant part is reproduced below:
"39.4 It has been accordingly contended that TPO committed many errors and was wrong in discrediting list of comparables found and furnished by the taxpayer. The TPO could have carried fresh search only if the comparables drawn by the taxpayer was insufficient or had other deficiency. We are of the view that objection raised on behalf of the taxpayer are well-founded and were wrongly disregarded by the T.P.O."
Conversely, the ITAT upheld the view that once the comparables selected by the taxpayer are found functionally dissimilar or otherwise unsuitable the TPO is justified in looking for fresh comparables. The TPO is thus duty bound to examine the comparables selected or rejected by the taxpayer. If a company is rejected by the taxpayer is found to be acceptable after FAR analysis and other comparability criteria, then the same would have to be included in the set of comparables.
(c) In para 46 of the order, the Hon'ble Tribunal referred to exclusion of high profit and high loss making companies as comparables. The point has not been clarified by the Hon'ble Tribunal as to what will constitute 'high profit' or 'high losses'. The TPO in the instant case tried to normalize the high margin cases. The taxpayer had selected comparables with margins varying from 2.15 per cent to 52.97 per cent.
(d) The ITAT held that turnover and size of comparable companies is a very pertinent criterion of comparability analysis. The TPO selected comparables which are closer in size and turnover to the taxpayer than the comparables selected by the TPO.
(e) The ITAT held that assets owned and employed by the comparable companies should also be considered in comparability analysis. The assets of the comparables selected by the taxpayer were grossly dissimilar to the taxpayer. Hence TPO rejected the same.
(f) The ITAT held that companies having related party transactions should not be taken as comparable. The TPO tried to follow thus and excluded such companies as per the data available at the time.
(g) The ITAT held that suitable adjustments should be allowed to the taxpayer for differences in risk etc. It must be pointed out that in the present case the plea for making adjustments on account of risk and working capital was not made at any time during the assessment proceedings and the claim was made before the CIT(A) for the first time. The taxpayer's claim of 5 per cent risk adjustment was without any proper basis. This has already been discussed above.
(h) The ITAT also gave some directions on functional comparability but the same are not very specific. In the Mentor Graphics (Nokia) (P.) Ltd.'s case ( supra), the ITAT finally accepted the taxpayer's comparable which included a trading company. Thus the ITAT's discussion on functional comparability actually does not lay down norms or guidelines on the subject.

4.73 In the present case the taxpayer as well as the TPO has selected comparables from a wide spectrum of software companies subject to the turnover criterion. The companies engaged in services other than software development were excluded. The software service providers were accepted as comparable irrespective of the verticals of the software development. The same approach had been followed by the taxpayer. Therefore the taxpayer cannot now find fault with the TPO following the Mentor Graphics (Nokia) (P.) Ltd.'s case (supra ), because on most of the issues the TPO has followed the line of the taxpayer himself.

4.74 In the para 46.2 of the order the Tribunal has considered the issue of adoption of mean PLI. The relevant Para is extracted below :

"46.2 While holding so, we have not adopted mean profit of several comparable found by respective parties because in spite of our repeated requests, the parties before us, were unable to show us any rule or decision under which average or mean margin (OP/TC) of different companies is to be taken. Tax administration and parties can work different Arm's length price i.e., a range by the application of different methods, in such a situation, mean of Arm's Length Price as provided in proviso to section 92C(2) of the Act can be taken. But above Arm's length range is not the same thing as average operating profits of different entities with different FAR worked through the same method as done in this case by adopting TNMM. The assessee has satisfied not one but several points of Arm's length range worked out on record. In our considered view, it is not necessary for the taxpayer to satisfy all points in the range. Even if one point is satisfied, the assessee can be taken to have established its case and in that situation, the onus is shifted to the department to show why taxpayer's case be not accepted. Arm's length price does not mean maximum price or maximum profit in the range. A willing buyer in an open market shall pay minimum and not maximum price for goods or services. Of course, quality and brand name are important but considered not so by T.P.O. as TNMM method was applied by him. Project profile and other factors were, therefore, not erroneously considered. As noted earlier, the case of Integrated Hitech has been specifically accepted as comparable by both the parties. On other four cases noted above, the T.P.O. or other revenue authorities have not made any adverse comment at any stage of proceeding. It was open to them in proceedings before the ld. CIT(A) or the Appellate Tribunal to show that PIL figure of Integrated Hitech or other four companies were wrong or on account of their FAR analysis, these entities could not be taken as "reliable" compatibles for computation of the Arm's Length Price. But no material was brought on record, no arguments advanced to reject the above transaction. Therefore, having regard to facts of the case and material on record, we accept them as comparable and accept the price disclosed by the taxpayer as Arm's Length Price. Consequently, the addition of Rs. 1,45,73,857 is directed to be deleted. The view taken by us finds support from para 1.4 of OECD guidelines which we quote below:—

'1.48 If the relevant conditions of the controlled transactions (e.g., price or margin) are within the arm's length range, no adjustment should be made. If the relevant conditions of the controlled transaction (e.g., price or margin) fall outside the arm's length range asserted by the tax administration, the taxpayer should have the opportunity to present arguments that the conditions of the transaction satisfy the arm's length principle, and that the arm's length range includes their results. If the taxpayer is unable to establish this fact, the tax administration must determine how to adjust the conditions of the controlled transaction taking into account the arm's length range. It could be argued that any point in the range nevertheless satisfies the arm's length principle.'"

4.75 The Tribunal appears to have given preference to the OECD guidelines over the provisions of the Indian Income-tax Act, which with due respect does not seem to be correct. In the case of Aztech Software & Technology Services Ltd. (supra), the Special Bench has observed as under at para 159 page 112 :

"A useful reference can always be made to OECD Guidelines, for the purposes of resolving dispute of transfer pricing in India, subject, however, to statutory regulations."

As per the provisions of section 92C(1):

"The arm's length price in relation to an international transaction shall be determined by any of the following methods, being the most appropriate method, having regard to the nature of transaction or class of transaction or class of associated persons or functions performed by such persons or such other relevant factors as the Board may prescribe, namely:

(a) comparable uncontrolled price method;
(b) re-sale price method;
(c) cost plus method;
(d) profit split method;
(e) transactional net margin method;
(f) such other method as may be prescribed by the Board."

From the above, it can be noted clearly that the first step in determination of the ALP is selection of the most appropriate method. The terms 'any of the following methods' and 'most appropriate method' have been used in singular which shows that there can only be one most appropriate method.

The statutory provisions of section 92C(2), in this regard are discussed below:

"The most appropriate method referred to in sub-section (1) shall be applied, for determination of arm's length price, in the manner as may be prescribed:

[Provided that where more than one price is determined by the most appropriate method, the arm's length price shall be taken to be the arithmetical mean of such prices, or, at the option of the taxpayer, a price which may vary from the arithmetical mean by an amount not exceeding five per cent of such arithmetical mean.]"

It is noteworthy that there can be only one most appropriate method because the section uses the term in singular.

Rule 10B(1) reproduced below also shows that there can only be one most appropriate method:

"For the purposes of sub-section (2) of section 92C, the arm's length price in relation to an international transaction shall be determined by any of the following methods, being the most appropriate method, . . . ."

Similarly heading of rule 10C is "Most appropriate method" not - methods. The rule also refers to most appropriate method in singular only.

Rule 10(D)(1)(i) also mentions only one most appropriate method selected by the taxpayer as under:

"a description of the methods considered for determining the arm's length price in relation to each international transaction or class of transaction, the method selected as the most appropriate method along with explanations as to why such method was so selected, and how such method was applied in each case"

In the statutory 3CEB Form also, the term used is 'most appropriate method' and not methods.

The aspect of the most appropriate method and average markup of comparable companies was discussed by the Hon'ble SC in its order in the case of Morgan Stanley & Co. (supra), where the court held as under:

"The taxpayer is required to compute arm's length price for a transaction(s) using one of the five methods stipulated in the Income-tax Rules. Rule 10C(1) of the Income-tax Rules defines the most appropriate method as the method which is best suited to the facts and circumstances of each particular international transaction. As per rule 10C(2) the most appropriate method has to be selected having regard to a number of factors which are enumerated therein:"

The Hon'ble Court further observed as under:

". . . . the methods quoted above namely CUP, RPM, CPM, PSM & TNMM are mentioned in section 92C read with rule 10B. The most appropriate method has to be applied for computation of the arm's length price. It will depend on facts and circumstances of each particular international transaction. . . ."

The Supreme Court has also held TNMM to be the most appropriate method in the case of Service PE —

"in our view apart from the orders passed by the TPO/Assessing Officer the said method (TNMM) is the appropriate method in the case of service PE as the TNMM apportions the total operating profit arising from the transactions on the basis of sales, costs, assets etc.,".

4.76 Thus the Tribunal's observation that tax administration and taxpayer can work different ALP's by application of different methods is not correct as per the law. Most appropriate method has to be one method only. The proviso to section 92C(2) regarding adoption of arithmetical mean is applicable if the same method throws up different prices. Method has to be selected first before working out the uncontrolled prices/PLI for comparability. The final ALP has to be one definite figure. There is no concept of an ALP range under the Indian Tax Provisions. The taxpayer's income from an international transaction has to bo determined with regard to the ALP so determined.

4.77 In the Mentor Graphics (Nokia) (P.) Ltd.'s case (supra) and also in the instant case, the taxpayers have correctly adopted the mean of the PLI as an ALP. If the taxpayer's profit were in divergence with the ALP an adjustment is required as per the law. The Tribunal view that no adjustment is required if the Taxpayer's PLI is comparable with any of the comparable company's PLI, with due respect, does not find support in the provisions of the Act. The Tribunal is correct in observing that the ALP does not mean maximum price or maximum profit in the range. But the Tribunal's finding that the ALP implies minimum price or profit in the range is not supported by the provisions of the Act. As per the provisions of the Act, the arithmetical mean of the different prices has to be taken as ALP. The Tribunal's observations are therefore contrary to the explicit provisions of the Act. It also appears that there was no controversy regarding adoption of arithmetical mean in the present case. This was not one of the grounds of appeal to be decided by the Tribunal. Before concluding it is also clarified that the ITAT decision in the case of Mentor Graphics (Nokia) (P.) Ltd. (supra) was essentially based on the peculiar facts of the case. The Hon'ble Tribunal allowed the taxpayer's appeal because of certain omissions and mistakes on the part of the TPO. This aspect is discussed at paras 34 to 41 of the ITAT order.

4.78 In brief the ITAT found that :

(a) The TPO erred in considering the data for the financial year 2003-04 while determining the related party transactions in the case of comparables of the financial year 2001-02.
(b) The TPO had adopted contradictory approaches by rejecting the taxpayer comparable for being not based on contemporaneous data while himself using data of the subsequent years for comparability.
(c) The TPO did not look into FAR analysis of the comparable companies.
(d) The TPO rejected 08 of the taxpayer comparables without making any specific comments.
(e) The TPO selected companies having vide range of profit margin.
(f) Certain inconsistencies were noticed in the TPO's selection/rejection, even after applying his own filters.

4.79 The peculiar facts and circumstances of the case of Mentor Graphics (Nokia) (P.) Ltd. (supra) are not present in the case of the taxpayer. Therefore, it will be improper to apply this decision in this case without first properly examining facts and circumstances of each case.

4.80 In the case of Aztec Software & Technology Services Ltd. (supra), it was held by the Spl. Bench that the ALP determined by the taxpayer should not be accepted merely because the ALP determined by the TPO staffers from certain defects. The Appellate Authorities should give specific findings before accepting the taxpayer's PLI.

"Having regard to the purpose of the legislation and application of similar enactment world over, it must further be held that adjustments made on account of ALP by tax authorities can be deleted in appeal only if the appellate authorities are satisfied and records a finding that ALP submitted by the assessee is fair and reasonable. Merely by finding faults with the transfer price determined by the revenue authorities (Assessing Officer/TPO), addition on account of "adjustments" cannot be deleted. This is because the mandate of section 92(1) is that in every case of international transaction, income has to be determined having regard to ALP. Therefore, unless ALP furnished by the taxpayer is specifically accepted, the appellate authorities on the basis of material available on record has to determine ALP itself. Subject to 'statutory' provisions, Appellate authorities can direct lower revenue authorities to carry this exercise in accordance with law. The matter cannot be left hanging in between. ALP of international transaction has to be determined in every case." (Page 130 supra)

4.81 In the Mentor Graphics (Nokia) (P.) Ltd.'s case (supra) the Tribunal accepted the PLI determined by the taxpayer irrespective of the fact that the TP clone by the taxpayer suffered from similar lacuna. The taxpayer's comparables also had huge variation in margins, the FAR analysis done by the taxpayer was faulty because it included even trading companies, the comparables selected by the taxpayer were also engaged in diversified activities and different verticals of software development, the taxpayer had also not looked into the related party transactions of the selected comparables and the taxpayer had also used earlier year data. Following the Aztec Software & Technology Services Ltd. order, with due respect the Tribunal should not have accepted the taxpayer's TP study but should have re-determined ALP afresh or should have directed the TPO to determine ALP afresh. It appears from the Tribunal's order that the Tribunal did not undertake a fresh TP analysis because the Tribunal was convinced that as one of the comparables had a PLI which was lower than the taxpayer's PLI hence there was no need for any adjustment or fresh TP analysis. It has already been discussed above at para 17.5 that the Tribunal's approach on the issue is not in line with the provisions of the Act. Had the Tribunal not followed this approach, it is apparent that the Tribunal would have undertaken a fresh TP analysis to determine the ALP.

4.82 The taxpayer is a part of the software industry in India. It is worth mentioning that NASSCOM survey of the software industry in India revealed that on an average industry operating margins were 26 per cent on sales i.e., 35.13 per cent on cost in the year 2005-06. The major part of the IT Industry in India consists of software exports i.e., 75 per cent where as ITES was only 25 per cent when exports are considered. An independent agency M/s. D&B had also done a research regarding cost structure and profitability of Indian software industry. As per D&B survey, the smaller software companies, having turnover less than Rs. 200 crore, had PBIT on sales at the rate of 18.4 per cent i.e., 22.50 per cent on cost in the year 2005-06.

4.83 Further the NASSCOM has also observed that based on the sample of 116 IT-ITES companies accounting for around 54 per cent of the export revenues as of March 2006 suggests that operating margins have shrunk from around 30 per cent in 2000-01 to 26.7 per cent in 2005-06. Thus there is empirical data available which suggests that the Indian software industry overall had very good profitability. The extremely low margins of the taxpayer seems to be an aberration and the burden is on the taxpayer to explain how and why it could not perform in line with the industry.

4.84 Before concluding, it is submitted that in case the Hon'ble ITAT decides that some more filters should have applied or certain filters should not have been applied or that the turnover range adopted by the TPO was not correct and a new range is to be adopted etc., the ITAT may kindly remand the matter back to the file of the Assessing Officer/TPO so that new comparables may be selected as per the ITAT guidelines. Any change in the search criterion will result into selection/rejection of additional comparables if a search of the data base is conducted as per the changed criterion. The changed criterion should not be applied only to reject the comparables already selected by the taxpayer/TPO. In this connection it is also mentioned here that even if it is ultimately held that the TPO's order suffers from some infirmities, it would not automatically establish that the taxpayer's transactions are at arm's length. The TP study done by the taxpayer has to stand on its own merits. Otherwise, the appellate authorities are required to determine the ALP of the impugned international transactions. In this regard the following observations made by the Special Bench of the ITAT in the case of Aztec Software Technology Services Ltd. (supra) are reproduced below:

"133. Having regard to the statutory provisions, particularly the mandate of sections 92(1) and 92D read with relevant rules, we hold that it is obligatory on the part of the taxpayer to furnish information relating to controlled international transactions, select a suitable method for determination and furnish ALP of such international transactions carried by it and give basis and supporting authentic evidence of ALP and adjustments made. The taxpayer has further to co-operate in the determination of the ALP by the tax authorities by furnishing all relevant information. The tax authorities in cases where they are of the opinion that ALP has not been correctly determined by the taxpayer, can substitute their own ALP on the basis of material or information furnished by the assessee or collected by them. However, such ALP has to be determined having in mind provisions of sections 92 and 92C and other rules and regulations. While determining ALP, tax authorities are bound to follow principles of natural justice and be fair and reasonable to the taxpayer. Any material collected to be used against the taxpayer is to be put to taxpayer to explain. Having regard to the purpose of the legislation and application of similar enactment world over, it must further be held that adjustments made on account of ALP by tax authorities can be deleted in appeal only if the appellate authorities are satisfied and records a finding that ALP submitted by the assessee is fair and reasonable. Merely by finding faults with the transfer price determined by the revenue authorities (Assessing Officer/TPO), addition on account of "adjustments" cannot be deleted. This is because the mandate of section 92(1) is that in every case of international transaction, Income has to be determined having regard to ALP. Therefore, unless ALP furnished by the taxpayer is specifically accepted, the appellate authorities on the basis of material available on record has to determine ALP itself, subject to statutory provisions, Appellate authorities can direct lower revenue authorities to carry this exercise in accordance with law. The matter cannot be left hanging in between. ALP of international transaction has to be determined in every case."

It has been amply shown in the above submissions that the TP study furnished by the taxpayer had so many infirmities. The taxpayer had selected comparables (Melstar and Softsol) who had more than 50 per cent related party transactions. The selection was made against the taxpayer's own filter of "predominant RP transactions". The taxpayer did not use contemporaneous data. The taxpayer had selected companies which were not comparable in size and turnover. The infirmities in taxpayer's TP study would not be cured simply because the TPO is alleged to have committed some mistakes. If the TPO's analysis is rejected on the grounds raised by the taxpayer, the taxpayer's analysis is also liable to be rejected on similar grounds. The Hon'ble Tribunal may accordingly have to determine the ALP of the taxpayer's international transactions or the matter may be restored to the Assessing Officer/TPO.

It will not be out of place here to bring to the notice of the Tribunal, certain inconsistencies in the approach and arguments of the taxpayer.

The taxpayer has not commented on the issue of the turnover filter applied by the TPO. It is admitted fact that but for M/s. Visual Soft, all other comparables selected by the taxpayer were very small in size when compared with the taxpayer. No specific arguments have been made by the taxpayer on this issue. If it is held that the lower limit of turnover (Rs. 100 crore) applied by the TPO was not justified, then the higher limit of the turnover applied by the TPO and the taxpayer (Rs. 250 crore) would also have to be relaxed. The entire search criterion would change and the TPO as well as the taxpayer should be allowed to search for the comparables within the new turnover limits if any determined by the Tribunal.

During the course of hearing before the Hon'ble Tribunal it was argued by the taxpayer that the turnover range of Rs. 100 crore to Rs. 250 crore would throw up more than 200 software companies. This appears to be factually incorrect. Even today, the databases do not show more than 20 to 30 software companies in this turnover range. However, if the taxpayer is correct, then the Hon'ble ITAT may direct the taxpayer to produce the list of the 200 software companies having turnover between Rs. 100 crore to Rs. 250 crore so that suitable comparables may be searched from these companies.

The taxpayer has argued against the 'significant' related party transactions filter, stating that the word 'significant' is TPO's innovation. The fact is that the taxpayer himself had applied a filter of rejecting companies having 'predominant' related party transactions (Vide Page 97 and Page 111 of the taxpayer's submissions). However the taxpayer did not clarify the meaning of 'predominant' in the TP study. The relevant portion of the TP Report (Page 111 of the taxpayer's submissions) is reproduced below.

"Step 9: Next, 8 companies were eliminated, as they were rendering predominant services to their associated enterprises having us with 16 companies for comparison."

Thus finding fault with the TPO without mentioning that the taxpayer himself had adopted similar approach shows the double standards being applied by the taxpayer.

4.85 The double standard followed by the taxpayer is also apparent from the fact that he is arguing selectively against high margin comparables selected by the TPO or the taxpayer himself like M/s. Visual Soft. In this regard, the following points are worth mentioning.

M/s. Visual Soft is a comparable selected by the taxpayer in its TP report. It is a software development company. It has income from export of software and it does not have any related party transactions either. The company is thus a perfect comparable selected by the taxpayer.

The taxpayer did not object to this company being taken as a comparable either before the TPO during the TP proceedings or before the CIT(A) in the appellate proceedings. The taxpayer did not object to this company as a comparable in the written submissions made before the Hon'ble Tribunal. However, as an afterthought, the taxpayer is now arguing before the Hon'ble Tribunal against this company and asking for its rejection. The taxpayer should not be allowed to take a stand which he did not take either before the TPO or before the CIT(A).

The taxpayer has put up an argument that because M/s. Visual Soft has incurred expenditure on R&D more than 3 per cent, it is not a comparable. In this regard, it is submitted that R&D expenditure was not used as a filter either by the taxpayer in its TP or by the TPO. The taxpayer did not raise this objection either before the TPO or before the CIT(A). The taxpayer should not be allowed to contradict its own TP study at this stage. It is also worth mentioning here that the taxpayer (PSCPL) himself has a full-fledged R&D division known as Philips Research as mentioned on Page 14 of the TP study (Page 39 of the taxpayer's paper book). The relevant portions of the TP Report are reproduced below to show that the taxpayer's main business is product related R&D:

"The professionals working - at PSCPL use state-of-the-art software engineering paradigms and platforms including real-time systems, components-based software engineering, multithreaded architecture and ASIC design methodologies in deep sub-micron technologies to drive the creation of tomorrow's products and services. PSCPL is an ISO 9001/Tick IT, SEI CMM SM Level 5 company and has emerged as a critical partner in the development of a range of Philips products... PSCPL has built up extensive know-how and expertise in the software engineering and technology-domains relevant to its business. PSCPL's software expertise is primarily in the areas of embedded and information system engineering, architecture design, programming and testing. PSCPL's IC design expertise is in the areas of logic and circuit design. In addition, competencies in the areas of project management, requirement engineering and quality assurance have been established to offer customers, products and services of the highest quality, at the fastest time-to-market and at the lowest cost of ownership. . . .

Philips Research (PR)

PR Group is called Design Technology for Trile S - Software, Silicon and System. The research activities at PR will align to the International Research Programme, embedded platform and software technology. The group will be responsible Hi supporting the Long-Range Technical Objectives, Integrated Systems Design and Embedded Software Technology. The group is organised into 3 dusters:

Testing of software components and systems; Components system design; System-level design methodologies.

Research work will be targeted at Product Divisions like PS, PCE and PMS."

As per the above, the research work of the taxpayer is targeted at product development like embedded software. Apparently, M/s. Visual Soft is functionally not different from the taxpayer.

*The taxpayer has also not explained any rationale for putting the limit of 3 per cent. The cut of limit of 3 per cent of R&D expenditure over sales is adopted by the taxpayer in a very arbitrary manner. There are no studies to show that companies incurring R&D expenditure exceeding 3 per cent of turnover are necessarily product companies, it may also be mentioned here that the nomenclature R&D expenditure in the P&L account necessarily does not indicate product related R&D. The salary expenditure etc., incurred on hiring of professionals is sometimes shown partly as R&D expenditure to show a better profile of the company. In most of such cases, the R&D will actually be the activities aimed at cost cutting and improving organizational setup etc. Therefore there is no reason to exclude comparable companies incurring some expenditure on R&D as long as they are overall software service providers.

*The taxpayer has argued that the R&D filter was upheld in the case of Mentor Graphics (Nokia) (P.) Ltd. (supra), Egain Communication (P.) Ltd.'s case (supra) and Development Consultants (P.) Ltd. v. Dy. CIT [2008] 23 SOT 455 (Kol.) . In this regard it is submitted that in none of the above decisions, the ITAT had actually decided the issue of suitability of R&D filter. The filter finds a mention in the ITAT orders simply because many taxpayers use this filter in search of the comparables.

*As held by the Special Bench of ITAT, Ahmedabad in the case of Asstt. CIT v. Goldmine Shares & Finance (P.) Ltd. [2008] 302 ITR 208 (AT) (SB), if an issue is not discussed and directly decided in a particular order, mere mention of the issue does not have any precedent value. Expenditure on R&D is not a suitable filter in the facts and circumstances of the case, as the taxpayer himself engaged in software R&D services. Thus, the taxpayer's case is clearly distinguishable on facts on the above-mentioned ITAT decisions relied by the taxpayer.

*The taxpayer has argued it should be allowed benefit of +/-5 per cent range as per the proviso to section 92C(2). It is submitted that the issue has been dealt with in detail by the CIT(A) in her order. The taxpayer's reliance on the Tribunal's decision in the case of Development Consultants (P.) Ltd. v. Dy. CIT [2008] 23 SOT 455 (Kol.) is of no avail, because in that case no arguments, were relied on this issue either by the Department or by the taxpayer. The computation of the ALP as submitted by the taxpayer before the Tribunal was not refuted by the Departmental Representative and hence was approved by the ITAT as under:

"22. The learned AR submitted that in case of assessment year 2003-04, the arm's length GP/Sales of the comparable companies is 28.22 per cent which is lower than the GP/Sales of DCIL for the year ended 31-3-2003 of 61.10 per cent indicating that DCIL has retained more than the arm's length margin at the gross level. He stated that transfer pricing legislation as provided in the IT Act allows a taxpayer to have an option to compute the ALP which may vary from the IT Act which states that "where more than one price is determined by the most appropriate method, the arm's length price shall be taken to be the arithmetical mean of such prices or at the option of the assessee, a price which may vary from the arithmetical mean by an amount not exceeding five per cent of such arithmetical mean". The assessee computed the arm's length, price considering the 5 per cent tolerance range. The results of such computation" are given below:

Shri Rahul Mitra submitted that based on the above analysis, the amount of adjustment in the international transaction of the assessee with DCIL would be USD 275,632. In case of assessment year 2004-05, the arm's length GP/Sales of the comparable companies is 28.75 per cent, which is higher than the GP/Sales of DCIL for the year ended 31-3-2004 of 25.26 per cent, indicating that DCIL has not retained more than the arm's length margin.

23. The ld. DR could not refute the arguments and facts discussed by Shri Rahul Mitra. He only reiterated that the Commissioner (Appeals) during the course of appellate proceedings had argued against the analysis that DCIL was the main company and that the assessee was only a paper company and not a company of substance. Based on his findings from the internet the Commissioner (Appeals), had understood the assessee to be the main company of the Group and DCIL, he argued was a paper company created only to evade taxes. Hence, in his order he stated that there is no justification for allowing 28 per cent margins to DCIL and entire profits should come to DCPL.

25. We have examined all the facts and record and after the due consideration of arguments placed by both the sides, we can see that DCIL has long standing banking relationship with banks like, Bank of India and Indian Overseas Bank. Further, it enjoys substantial credit limits provided by these banks for the purpose of its business operations. No paper company would be having such substantial business operations and have excellent customer relationship with its bankers and no nationalized banks would give such credit limits to a paper company. DCIL has been established after proper permission obtained from the Reserve Bank of India. Hence, based on the facts and our findings we can conclude that DCIL is a company of substance and is performing full-fledged distribution activities. It is not a paper company established to evade taxes as argued by the Commissioner (Appeals) in his order. Hence, we do not find any justification in the arguments of the Commissioner (Appeals) that entire profits should come to the assessee and DCIL should not retain profits. Further, the benchmarking exercise and analysis conducted by the assessee has been examined by the Commissioner (Appeals) and the TPO, but they have not been able to controvert the analysis of the assessee. Therefore, we conclude DCIL should retain the gross margins as determined through the benchmarking exercise by the assessee discussed earlier in this order. Therefore, based on such analysis, in case of assessment year 2003-04, the amount of adjustment in the ALP of the international transactions entered by the assessee with DCIL should be restricted to USD 275,632 as submitted and conceded by the learned AR. This amount converted to Indian rupees based on the average currency conversion rates for the relevant year comes to Rs. 1,33,42,185. In case of assessment year 2004-05, based on the analysis discussed above, the assessee has conducted its international transactions with DCIL at arm's length hence there should not be any adjustment in the arm's length price for this year."

As can be seen from the above, the Departmental Representative did not refute the computation given by the taxpayer after considering +/-5 per cent range. Further, the 'Tribunal did not discuss the issue of +/-5 per cent in the above part of the order. Thus it does not amount to deciding all the issues involved in the computation of ALP including +/-5 per cent Range. Reference is gain invited to the decision of ITAT, Special Bench in the case of Goldmine Shares & Finance Ltd. (supra).

5.1 We have heard the rival contentions and we proceed to adjudicate on the issues in the sequence which has been argued by the rival parties before us. The learned counsel for the assessee has argued that the tax payable by it in India is lower than the tax rate applicable to its associated enterprise in the Netherlands. Since the assessee is availing the benefit under section 10A of the Act, one cannot take a simplistic view on the matter of tax avoidance. In this connection the learned Departmental Representative has drawn reference to the proviso to section 92C(4). Relying on OECD guidelines, the DR has mentioned that the consideration of transfer pricing should not be confused with the consideration of problems of tax avoidance, even though transfer pricing policies may be used for such purposes. In this connection, it was pointed out that by not declaring proper profits in India, the assessee is indirectly reducing its liability to Dividend Distribution Tax (in short 'DDT'). The Special Bench of the Tribunal, in the case of Aztec Software & Technology Services Ltd. (supra), has concluded that the Assessing Officer/TPO need not prove the motive of shifting of profits outside India for making a transfer pricing adjustment. The assessee had generally argued that one of the factors driving any motive for shifting profits would be the difference in the tax rate in India and the tax rate applicable to the associated enterprise in the overseas jurisdiction. In the instant case, since the assessee was availing the benefit under section 10A of the Act, it would be devoid of logic to argue that the assessee had manipulated prices (and shifted profits) to an overseas jurisdiction for the purpose of avoiding tax in India. The reference by the DR to the proviso to section 92C(4) is completely out of context and irrelevant. The DR ought to have appreciated that the proviso comes into play only once a transfer pricing adjustment is made. By quoting OECD guidelines, the ld. DR does not get much help. The ld. DR ought to have appreciated that what is relevant in the Indian context are the specific provisions of the Circular No. 14/2001. As submitted above, at para 55.5 of the said Circular, the CBDT has clearly mentioned that the intention of the transfer pricing provisions is to curtail avoidance of taxes by shifting profits outside India.

5.2 Further, the learned DR's argument on DDT is not on solid ground and has no bearing on any motive of shifting profits. This is clear from the fact that it is not uncommon for companies to use retained earning for expansion in India. The arguments of the DR on motive for avoiding DDT would fail in such a scenario. While making the argument the DR has drawn an incomplete reference to the decision of the Tribunal, in the case of Aztec Software & Technology Services Ltd. (supra). The DR has drawn limited reference to para 17 of the said order, to make an argument that there is no need to demonstrate a motive of tax avoidance for making a transfer pricing adjustment, however, as earlier submitted by the assessee, if one were to read the whole of the para 17, it is very clear that the Tribunal's observation was limited to the matter that no tax avoidance motive needs to be demonstrated at the time of initiation of the transfer pricing proceedings. The Tribunal has not discussed as to whether the TPO/Assessing Officer needs to demonstrate such a motive at the time of framing the assessment. Para 17 of the said order has been reproduced below :

"17. In view of the above judgments, the CIT(A) was not justified in holding that avoidance of tax is a condition precedent for invoking the provisions of section 92C/92CA of the Act. He was also not justified in attaching too much importance to Budget Speech. Notes on clauses, Memorandum to Finance Bills, apart from heading of the Chapter. The assessee can further derive no support from provision of section 52 or decision of the Hon'ble Supreme Court in the case of K.P. Varghese vs. ITO (supra) ( 131 ITR 597) = 2002-TIOL-128-SC-IT wherein profit on account of understatement of consideration was held to be essential special requirement of the provision. There is no such requirement of establishment of "tax evasion" before initiation of proceedings for determination of arm's length price as discussed above. Contrary view held by the learned CIT(A) is accordingly held to be unsustainable."

5.3 Further, the DR's argument on this point is also not consistent with the provisions of section 92C(3) of the Act and the Circular No. 14/2001 which provide that before the arm's length price is determined by the Assessing Officer, he has to prove that at least one of the four conditions laid down in sub-section (3) above have been satisfied. In the instant case, the Assessing Officer did not prove to the assessee that the above conditions were satisfied, either before initiating the transfer pricing assessment or during the course of the proceedings.

5.4 Circular No. 14/2001 reported in 252 ITR 65 (Cir.) (2001)issued by the CBDT spells out the intention of inserting the provisions of transfer pricing in the Act, vide Finance Act, 2001. Relevant extracts of the CBDT Circular is reproduced as under :

"55.5 The new provision is intended to ensure that profits taxable in India are not understated (or losses are not overstated) by declaring lower receipts or higher outgoings than those which would have been declared by persons entering into similar transactions with unrelated parties in the same or similar circumstances. The basic intention underlying the new transfer pricing regulations is to prevent shifting out of profits by manipulating prices charged or paid in international transactions, thereby eroding the country's tax base."

5.5 As per the above mentioned circular, the transfer pricing provisions ought to be applied to cases where shifting of profits happens outside India by manipulation of prices. In the instant case, the TPO/Assessing Officer did not establish, either before initiating the transfer pricing proceedings or even at the time of concluding the proceedings, that the assessee had manipulated prices to shift profits outside India. In this connection, the learned counsel drew our attention to Instruction No. 3, dated 20-5-2003 reported in 261 ITR (St.) 51 issued by the CBDT, which makes it mandatory for an Assessing Officer to make a reference to the TPO under section 92CA of the Act, in case the aggregate value of the international transactions of an assessee exceeds Rs. 5 crores. Relevant extracts of the said instruction has been reproduced below :

"(i )Reference to Transfer Pricing Officer (TPO)..... The Central Board of Direct Taxes, therefore, have decided that wherever the aggregate value of international transaction exceeds Rs. 5 crores, the case should be picked up for scrutiny and reference under section 92CA be made to the TPO. If there are more than one transaction with an associated enterprise or there are transactions with more than one associated enterprises the aggregate value of which exceeds Rs. 5 crores, the transactions should be referred to the TPO."

It is well-settled principle that instructions/circulars issued by the CBDT have a binding effect on the Assessing Officer. More specifically, the binding nature of the above-mentioned instruction issued by the CBDT has been confirmed by the Hon'ble Delhi High Court in case of Sony India (P.) Ltd. v. CBDT [2007] 288 ITR 52 . At para 37 of the said order, the Delhi High Court has concluded as below:

"For these reasons, we hold that the impugned Instruction No. 3, dated 20-5-2003 issued by the CBDT is consistent with the statutory objective underlying section 92CA(1) and acts as a guidance to the Assessing Officer in the exercise of discretion in referring an international transaction to the TPO for determination of its ALP. It is neither arbitrary nor unreasonable, and is not ultra vires the Act."

This view has also been taken by the ITAT, Bangalore Bench in the case of Aztec Software & Technology Services Ltd. (supra). At paras 44 and 45 of the order, the Tribunal has observed as below:

"44 . . . . Therefore, when provisions of section 92CA(1) are read with Circular of CBDT (Instruction No. 3), dated 20-5-2003, it becomes 'necessary' for the Assessing Officer to refer the question of determination of transfer pricing of international transactions to the TPO. He has no discretion in the matter in the light of limit fixed by the board. The Assessing Officer has only to look at the aggregate value of international transactions disclosed by the assessee in the audit report and then follow directions of CBDT. The Assessing Officer, therefore, is left with a very limited role under section 92CA(1). Likewise while granting approval to the action of the Assessing Officer the CIT has only to see whether aggregate value of international transaction is more than Rs. 5 crores or not. If it is more, he has also to grant approval in the light of directions of CBDT. These directions are mandatory and binding on the Assessing Officer and the CIT(A). The fact that directions are not binding on the assessee or courts is immaterial. The relevant question to be answered is whether circulars are binding on the departmental authorities. The question has an obvious answer.

The issue of legality and validity of above directions was raised before Hon'ble Delhi High Court in the case of Sony India (P.) Ltd. (supra). Their Lordships have also upheld validity of directions of CBDT.

45. In the light of above discussions, we do not find any illegality in the directions issued by the Board of Direct Taxes (CBDT)."

Since the circulars issued by the CBDT have a binding affect on the tax authorities, Circular No. 14/2001 issued by the CBDT would also have a binding effect on the Assessing Officer/TPO who is duty bound to demonstrate that the assessee has manipulated its prices to shift profits outside India, before any transfer pricing adjustment is made. Section 92C(3) provides that the Assessing Officer may proceed to determine the ALP if he is of the opinion that certain conditions have been satisfied. Relevant extracts of the said sub-section are reproduced as under:

"(3) Where during the course of any proceeding for the assessment of income, the Assessing Officer is, on the basis of material or information or document in his possession, of the opinion that -

(g )the price charged or paid in an international transaction has not been determined in accordance with sub-sections (1) and (2); or

(h)any information and document relating to an international transaction have not been kept and maintained by the assessee in accordance with the provisions contained in sub-section (1) of section 92D and the rules made in this behalf; or

(i )the information or data used in computation of the arm's length price is not reliable or correct; or

(j )the assessee has failed to furnish, within the specified time, any information or document which he was required to furnish by a notice issued under sub-section (3) of section 92D.

The Assessing Officer may proceed to determine the ALP in relation to the said international transaction in accordance with sub-sections (1) and (2), on the basis of such material or information or document available with him. . . ."

Hence, before the ALP is determined by the Assessing Officer, he has to prove that at least one of the four conditions laid down in sub-section (3) above have been satisfied. However, in the instant case, the Assessing Officer has not proved that the above conditions were satisfied, either before initiating the transfer pricing assessment or before the completion of the assessment proceedings. The learned counsel submitted that in the case of assessee, the Assessing Officer has failed to establish that any of these four conditions have been satisfied. In fact, there is not even a finding to this effect that any of the four conditions have been satisfied. Further, section 92CA(3) provides that even a TPO should determine the ALP in accordance with the provisions of section 92C(3). Accordingly, the conditions of section 92C(3) would also be relevant to the TPO. The learned counsel drew reference to Circular No. 2/2001, dated 23-8-2001, issued by the CBDT, wherein the following has been mentioned :

"(iii)It should be made clear to the concerned Assessing Officers that where an international transaction has been put to a scrutiny, the Assessing Officer can have recourse to sub-section (3) of section 92C only under the circumstances enumerated in clauses (a) to (d) of that sub-section and in the event of material information or document in his possession on the basis of which an opinion can be formed that any such circumstance exists. In all other cases, the value of the international transaction should be accepted without further scrutiny."

Based on the abovementioned para, it is clear that the intention of section 92C(3) has always been that scrutiny of the international transactions of an assessee can only be done if the Assessing Officer/TPO can prove that the circumstances enumerated in clauses (a) to (d) are satisfied. Even where any infirmity is identified by the Assessing Officer/TPO, the action of the Assessing Officer/TPO would be restricted to taking remedial action commensurate with the infirmity identified by him, and not beyond. For instance, if there is a finding, based on evidence, for satisfaction of the condition of section 92C(3)(d), the Assessing Officer/TPO could, at best, use his judgment as regards any information/document, unreasonably withheld by the taxpayer, for the purpose of making the assessment, on the other hand, for a case where condition of section 92C(3)(a) is triggering, and not triggering any of the other conditions of section 92C(3), the Assessing Officer/TPO has to use the data used by the taxpayer and modify the analysis of the taxpayer only to the extent that the computation of the ALP deviates from sub-sections (1) and (2) of section 92C.

Comparison with old section 145 - rejection of accounts.

A parallel on this provision can also be drawn to the provisions of section 145 as it stood prior to amendment with effect from 1-4-1997. Under the said provisions, an Assessing Officer had the power to make an assessment in the manner laid down in section 144, where he was not satisfied about the correctness or the completeness of the accounts of the assessee, or where no method of accounting was regularly employed by the assessee. The judicial position has been that accounts regularly maintained in the course of the business have to be taken as correct, unless there are strong and sufficient reasons that they are unreliable. The department has to also prove satisfactorily that account books are unreliable, incorrect or complete, before it can reject the accounts. Further, it has been judicially held that the Assessing Officer cannot reject the books of account, merely on account of low profits. We refer to the following two decisions in this context :

(i) Ramachandra Ramniwas v. State of Orissa [1970] 25 STC 501 (Ori.).
(ii) International Forest Co. v. Commissioner Of Income-Tax, Patiala. [1975] 101 ITR 721 (J&K).

In the instant case, the TPO/Assessing Officer have at no stage of the assessment proceedings established that the transfer pricing record of the assessee should be rejected. The TPO/Assessing Officer simply conducted their analysis and applied the same to the assessee's case. Once the assessee has discharged its onus by conducting a proper comparability analysis it would be open for the Assessing Officer to make a fresh analysis only if the case of the assessee falls under section 92C(3)(c). There is nothing in the order of the TPO or the Assessing Officer to suggest that clause (c) of section 92C(3) is satisfied. In this connection, reference is also drawn to Circular No. 14/2001 issued by the CBDT, which also provide that only if such deficiency or insufficiency is found in the documentation of the assessee, the TPO/Assessing Officer can conduct a scrutiny. The relevant extract of the Circular has been reproduced below :

"Where such onus is discharged by the assessee and the data used for determining the ALP is reliable and correct, there can be no intervention by the Assessing Officer. This is made clear by sub-section (3) of section 92C which provides that the Assessing Officer may intervene only if he is, on the basis of material or information or document in his possession, of the opinion that the price charged in the international transaction has not been determined in accordance with sub-sections (1) and (2), or information and documents relating to the international transaction have not been kept and maintained by the assessee in accordance with the provisions contained in sub-section (1) of section 92D and the rules made thereunder; or the information or data used in computation of the ALP is not reliable or correct. ..."

Further in the case of Mentor Graphics (Nokia) (P.) Ltd. (supra), the Tribunal, at para 39.4 of the order have observed as follows :

"The TPO could have carried fresh search only of the comparables drawn by the taxpayer was insufficient or had other deficiency."

The Act and the Rules provide that while conducting the comparability analysis, the data to be used should be contemporaneous. In this regard, the requirement of law is two-fold :

(a) Data to be used for analyzing the comparability of an uncontrolled transaction shall be the data relating to the financial year in which the international transaction has been entered into [Rule 10B(4)]; and
(b) Amongst other things, the data which is used for the comparability analysis should exist latest by the specified date [Rule 10D(4)].

It is important to note that rule 10B(4) casts an obligation on the taxpayer to conduct the comparability analysis using data for the relevant financial year. However, rule 10D(4) makes it mandatory for the taxpayer to ensure data that exists by the time specified under the Act, i.e., 31st October of the relevant assessment year. Further, rule 10C(2)(c) also clarifies that availability of data is a significant factor in conducting a comparability analysis, whether it relates to selection of the most appropriate method or arriving at a set of comparables or computing the margins of such comparables. Accordingly, for the purpose of conducting the comparability analysis, the data should :

(a) relate to the relevant financial year [if the proviso to rule 10B(4) is not attracted];
(b) exist by the specified date.

It should be noted that both the conditions are cumulative in nature. If any one of the conditions are not satisfied, the relevant comparable ought not to be included in the comparability analysis. In other words, if data for a particular comparable company for financial year 2002-03 was not available as on 30-11-2003 (i.e., the specified date for financial year 2002-03), the same cannot be used for conducting the comparability analysis. The transfer pricing analysis was conducted by the assessee in October, 2003 using Capitaline Plus database updated as on 30-9-2003, which is a period reasonably close to the specified date of 30-11-2003. However, the TPO conducted a fresh comparability analysis in the show-cause notice issued on 3-2-2006. The fact that the TPO conducted the comparability analysis after the 'specified date', clearly proves that such comparability analysis did not use data which was 'contemporaneous' and hence is not in compliance with the provisions of rule 10D(4).

5.6 If the provisions of rules 10B(4) and 10D(4) were to be applied to a book e.g., written in 1857 would satisfy all the conditions as - (a) it relates to the war of independence; (b) it is contemporaneous; and (c) it exists by the specified date (i.e., 31-12-1860). For that matter, even the book written in 1860 would satisfy the test as it would exist by the specified date and relates to the war of independence. However, the book written in 2008 would not satisfy the test as it would not be contemporaneous, and also would not exist by the specified date.

5.7 In the context of the argument of the assessee, the CIT(A) in her order (at pages 26 to 31 of her order) has observed that the provisions of rule 10D(4) do not override the provisions of rule 10B(4). Thus, inter alia, the data for financial year 2002-03 shall be used, even though the same was not available by the specified date. For the reasons as discussed above, the assessee submitted that the above observation of the CIT(A) is clearly contrary to the provisions of rules 10B(4) and 10D(4).The assessee therefore submitted that the TPO/CIT(A) is not justified in :

(a) conducting a fresh comparability analysis during the course of the assessment (also see our submissions under Issue Nos. 5, 6 and 12); and
(b) using non-contemporaneous data, i.e., data which was not available as on the specified date.

It would not be out of place to quote the following observation of the OECD on this matter :

"Each taxpayer should endeavour to determine transfer prices for tax purposes in accordance with the arm's length principle, based upon information reasonably available at the time of the determination."

The learned counsel for the assessee does not deny that as per rule 10B(4) and proviso thereof, data relating to the previous two financial years may be used, provided such data has an influence on computing the arm's length price. The reference by the DR to the decisions of Aztec Software & Technology Services Ltd. (supra) and Mentor Graphics (Nokia) (P.) Ltd. (supra) is not in the correct context as pointed. In any case, the learned counsel submits that the issue of use of multiple year data was not argued before the Tribunal in both the cases as cited by the DR, and that the observations of the Tribunal in both the said cases are only on the limited issue of rule 10B(4). The orders do not disclose any argument on the point raised by the learned counsel for the assessee. The DR has confused the issue relating to (a) use of relevant financial year's data, and (b) use of multiple year data i.e., use of data relating to more than one year - the case that is covered under the proviso to rule 10B(4). The learned counsel had neither in its transfer pricing study, nor during the course of transfer pricing assessment claimed to use multiple year data i.e., data relating to more than one financial year. In fact, while conducting the transfer pricing study, the learned counsel had clearly mentioned (refer page 95 of the paper book) that data relating to the year ended March, 2003 has been used and in case such data is not available, the previous year's data has been used. The process adopted in the TP study is clearly in conformity with the provisions of rules 10B(4) and 10D(4) which provide as follows :

(i) Rule 10B(4) - data relating to the relevant financial year shall be used.
(ii) Rule 10D(4) - data used should be as far as possible be contempo-raneous and such data should exist as on the specified date.

The ld. DR has made an argument that rules 10B(4) and 10D(4) have to be read independently and in isolation, and both the rules are for different purposes - rule 10B(4) is relevant for the purpose of comparability, while rule 10D(4) is relevant for avoiding penal provisions. The learned counsel submitted that such an interpretation of the ld. DR is completely erroneous and without any legal basis. If one were to take the argument of the ld. DR to a logical conclusion, a taxpayer would be required to maintain two separate sets of documentation - one that would satisfy the conditions of Rule 10B(4) and the other that would satisfy the conditions of rule 10D(4).

5.8 Rule 10B has been framed under section 92C and rule 10D has been framed under section 92D. Section 92C expressly refers to the documentation to in section 92D. Section 92D provides as follows :

"(1) Every person who has entered into an international transaction shall keep and maintain such information and documentation, in respect thereof, as may be prescribed. ..."

Further, the operative portion of rule 10B reads as follows :

"(1) Every person who has entered into an international transaction, shall keep maintain the following information and documents namely. ..."

On a combined reading of sections 92C and 92D, and rules 10B and 10D, it is clear that the data used for the purpose of conducting a comparability analysis should relate to the relevant financial year [if the proviso to rule 10B(4) is not attracted]; and be available as on the specified date. It should be noted that both the conditions are cumulative in nature. If any one of the conditions are not satisfied, the relevant comparable ought not to be included in the comparability analysis.

5.9 Another example that, the DR's argument that the provisions of rule 10D(4) does not have a bearing on rule 10B(4) is completely erroneous is that clause (g) of rule 10D(1) relates to records of the comparability analysis, as mentioned in rule 10B. The DR's interpretation gives a preference to rule 10B(4) over rule 10D(4) without any legal basis or justification. The DR should have appreciated that both the rules co-exist and should be read harmoniously. With regard to the above, it would not be out of context to mention the provisions of rule 10C(2)(c), which relates to selection of the most appropriate method. Relevant extracts of the sub-rule is reproduced as under :

"(c )the availability, coverage and reliability of data necessary for application of the method."

The above sub-rule also clarifies that availability of data is a significant factor in conducting a comparability analysis, whether it relates to selection of the most appropriate method or arriving at a set of comparables or computing the margins of such comparables.

5.10 Section 92C read with rule 10C provides that the arm's length price of an international transaction shall be determined by any of the five prescribed methods, being the most appropriate method. In connection with the above, while conducting the Transfer Pricing Study (TP Study), the learned counsel had selected the Cost Plus Method (CPM) as the most appropriate method, after evaluating the criterion laid down in Rule 10C(2), which are summarized as under :

(a) Functions performed, taking into account the risks assumed and the assets employed by the assessee;
(b) Availability and reliability of data;
(c) Degree of comparability; and
(d) Extent of reliable and accurate adjustments that can be made to account for differences.

Further, the guidance issued by the Institute of Chartered Accountants of India (ICAI), relating to issue of Form 3CEB by Chartered Accountants, explains the use of the most appropriate method as under :

"13.5 The conditions referred to above are cumulative. The reference there into the terms 'best suited' and 'most reliable measure' indicates that the most appropriate method will have to be selected after a meticulous appraisal of the facts and circumstances of the international transaction."

The guidance issued by the ICAI clearly specifies that selection of the most appropriate method should be based on a meticulous appraisal of facts and circumstances. In the TP Study, the assessee had carefully evaluated the five prescribed methods, and based on the facts, the assessee selected CPM as the most appropriate method. The analysis of the assessee is documented in pages 85 & 86 of the paper book. To support the results achieved by applying the CPM, the assessee had also selected the Transactional Net Margin Method (in short 'TNMM') in its TP Study. However, while conducting his comparability analysis and passing the order, the TPO did not provide any reasons for rejecting the CPM, which was selected as the most appropriate method by the assessee. Further, the TPO selected the TNMM as the most appropriate method, without sharing with the assessee any analysis, basis or reasons which led him to :

(a) select the TNMM as the most appropriate method; and
(b) reject CPM, as selected by the assessee, as the most appropriate method.

Sections 92C and 92E cast a responsibility on the assessee to prove that the requisite transfer pricing documentation, including the comparability analysis has been conducted and maintained by the assessee. Likewise, where a transfer pricing assessment is conducted by the TPO/Assessing Officer under section 92CA or 92C(3), it is the duty of the TPO/Assessing Officer to prove that such documentation maintained by the assessee is deficient or insufficient in any manner. Only if such deficiency or insufficiency is found in the documentation, the TPO/Assessing Officer can conduct a scrutiny. In this regard, relevant extracts of the CBDT Circular No. 14/2001 (para 55, 11) are produced below :

"... Where such onus is discharged by the assessee and the data used for determining the arm's length price is reliable and correct, there can be no intervention by the Assessing Officer. This is made clear by sub-section (3) of section 92C which provides that the Assessing Officer may intervene only if he is, on the basis of material or information or document in his possession, of the opinion that the price charged in the international transaction has not been determined in accordance with sub-sections (1) and (2), or information and documents relating to the international transaction have not been kept and maintained by the assessee in accordance with the provisions contained in sub-section (1) of section 92D and the rules made thereunder, or the information or data used in computation of the arm's length price is not reliable or correct. ..."

Further, it would be relevant to note that Instruction No. 3, dated 20-5-2003 (supra), also provides as under :

"(ii)Role of Transfer Pricing Officer

... The Transfer Price has to be determined by the TPO in terms of section 92C. The price has to be determined by any one of the methods stipulated in sub-section (1) of section 92C and by applying the most appropriate method referred to in sub-section (2) thereof."

As can be seen from the above, the CBDT has held that where :

- the onus is discharged by the assessee; and
- the data used by the assessee is reliable and correct, there can be no intervention by the Assessing Officer.

In the instant case, the assessee discharged its liability, inter alia, of

(a) preparing a transfer pricing documentation;
(b) conducting a comparability analysis, and
(c) furnishing the same to the TPO.

However, during the transfer pricing assessment proceedings, the TPO, without providing any reasons, followed a different methodology and selected a different method, completely disregarding the statutory provisions and the fact that :

- the assessee had discharged the onus of maintaining the prescribed documentation; and
- the data used by the assessee was reliable and correct.

Very clearly, in the instant case, the TPO/Assessing Officer has not discharged the onus of proving that the methodology followed by the assessee is erroneous in any manner, before rejecting the same. The learned counsel submitted that TPO is not justified in considering another method as the most appropriate method, even though there is no shortcoming in the method adopted by the assessee in the TP Study. Reference in this connection can also be made to the decision of the Delhi Tribunal in case of Mentor Graphics (Nokia) (P.) Ltd. (supra). At para 39.4 of the said order, the Tribunal has observed as follows :

"The TPO could have carried fresh search only of the comparables drawn by the taxpayer was insufficient or had other deficiency."

Further reference in this connection could also be made in the Madras High Court in case of CIT v. K. Sankarpandia Asari & Sons [1981] 130 ITR 541 .

5.11 A look at the show-cause notice referred to by the DR clearly shows that the argument of the DR is erroneous as in the said notice, no reasons were given for rejecting CPM as the most appropriate method (refer pages 231 to 235 of the paper book). Further, neither were any reasons set out in the order of the TPO and nor has the DR suggested that the reasons have been given elsewhere.

5.12 In the TP Study, the assessee had carefully evaluated the five prescribed methods, and based on the facts, the assessee selected CPM as the most appropriate method. The analysis of the assessee is documented in pages 85 and 86 of the paper book. The DR has referred to the observation of the Supreme Court, out of context :

- The Supreme Court's observation on TNMM being the most appropriate method, as referred to by the DR, was in the context of a service PE, and that too a passing comment. Moreover, the decision was rendered by the Supreme Court in the context of the specific facts of that case.
- The onus of determination of the most appropriate method is on the TPO and not that of the DR. The DR cannot discharge the functions of a TPO.
- Further, any conclusion can only be reached after comparing all the five methods, which the DR has not done. On the other hand, the assessee has carefully evaluated all the five methods and selected the CPM as the most appropriate method.
- The DR's argument of applying TNMM in all cases would mean that all the four other methods would be left redundant.

Accordingly, the said observation cannot be applied as such in the instant case, 11 would not be out of place here to mention that at para 27 of the said order, the Supreme Court has itself observed that :

"27. The methods, quoted above, namely, CUPM, MA;. Ravindra Paper Mart, CPM, PSM, TNMM etc., are mentioned in section 92C read with rule 10B. The most appropriate method has to be applied for computation of the arm's length price. It will depend on the facts and circumstances of each particular international transaction (see rule 10C)."

In the TP Study conducted by the assessee, the database used for conducting the comparability analysis was Capitaline 2000 ("Capitaline"). The said database is compiled by Capital Market Publishers India Limited and is a comprehensive interactive database of around 7,000 Indian companies, covering all companies listed or major stock exchanges like BSE/NSE plus other big unlisted companies. However, for the purpose of concluding the transfer pricing assessment, the TPO used another database (i.e., Prowess). The TPO did not :

(a) question the database used by the assessee;
(b) question the data which emanated from such database;
(c) specifically reject the database used by the assessee; and
(d) provide any reason for using the new database.

5.13 The learned counsel therefore submitted that since the TPO has not rejected the basis (including the database used) of conducting the comparability analysis by the assessee, using a completely different basis is an erroneous approach followed by the TPO. In other words :

(a) where the assessee has followed a methodical process while conducting the TP study, with a recognized database and a correct approach.
(b) the approach followed by the assessee is in consonance and in conformity with the requirements of the law; and
(c) there is no shortcoming or deficiency in the approach of the assessee. There is no reason as to why the TPO/Assessing Officer should conduct a fresh transfer pricing analysis, completely disregarding the approach of the assessee which is acceptable in law. The assessee submits that the approach followed by the TPO/CIT(A) to reject the TP Study of the assessee and to select a set of new companies is against the provisions of the Act/Rules and is based on their whims and fancies.

5.14 The learned counsel had submitted that if a methodology/basis/assumption/database adopted by the assessee is not found unacceptable by the TPO, the TPO cannot change the same merely on a whim or because he prefers otherwise. In this regard, reference could also be made in the Madras High Court in case of K. Sankarapandia Asari & Sons (supra).

5.15 The learned counsel reiterates that for the TPO ought to have—

(a) questioned the database used by the assessee.
(b) questioned the data which emanated from such database.
(c) specifically rejected the database used by the assessees, and
(d) provided reason for using the new database.

In theory, the comparability analysis in a transfer pricing documentation can be conducted either by cherry picking companies and considering them as comparables or starting with a set of all companies which are potentially comparables and following a methodical approach for eliminating non-comparable companies to leave a final set of comparables, consistent with the criteria used for elimination.

The assessee has conducted a TP Study using the second approach as discussed above. The TPO has resorted to pick and choose.

5.16 In the TP Study conducted by the assessee, comparable companies were arrived at after using a methodical search process on the Capitaline database. For the purpose of conducting the comparability analysis on Capitaline, the assessee selected all companies in the 'computer software' industry, as the first step. The said search process was followed by applying a number of pre-defined filters, both quantitative (i.e., system based) and qualitative (i.e., manual based) filters/eliminations. Thus, the final comparable companies were those which survived the elimination process and not the companies which were selected by the assessee. The final set of comparable companies were in accordance with the criterion mentioned in rule 10B(2). Rule 10B(2), provides that the following factors need to be considered for the purpose comparability analysis :

(a) the specific characteristics of the property transferred or services provided in the transactions;
(b) the functions performed, taking into account the assets employed and risks assumed by the respective parties to the transactions;
(c) the contractual terms of the transactions, dividing the risks, responsibilities and benefits between the parties to the transactions;
(d) conditions prevailing in the market of the respective parties to the transactions, including the laws, economic development, level of competition, etc.

It would also be relevant to note that the Transfer pricing guidelines for Multinational Enterprises and Tax Administrations issued by the Organisation for Economic Cooperation and Development ('OECD'), at pages 1-9 to 1-15, lays down the following five factors which should be considered for conducting a comparability analysis. The said five factors, which are very similar to the provisions of rule 10B(2), are as under :

(i) Characteristics of property or services;
(ii) Functional analysis;
(iii) Contractual terms;
(iv) Economic circumstances; and
(v) Business strategies.

From the above, it is clear that the focus is on the functions performed, and the reference to other economic criterion is only in the context of the functions. It would also be relevant to note that in the order (at page 251 of the paper book), the TPO has admitted that the comparables are functionally similar. The relevant extract is reproduced below :

"Even though the comparables selected in the TP document are functionally similar, their cases used in the business are grossly dissimilar to those of PSCPL. ..."

However, even after admitting that the compatibles in the TP Study are functionally similar to the assessee, the TPO has rejected the comparables in the TP Study Quantitative filters of turnover and asset base.

5.17 The learned counsel had argued that the TPO has erred in rejecting the comparables in the TP Study, merely on the basis of comparison of the asset base and the turnover of the companies, completely ignoring the fact that the comparables were the ones which survived a structured filter/elimination process, and the comparables are admittedly functionally similar to the assessee. A mere reference to assets employed or the turnover/size of operations of the comparables, disregarding the functions performed is an erroneous approach by the TPO/Assessing Officer.

5.18 The same can also be demonstrated by way of an illustration in case of a, say steel manufacturing company. If the steel manufactured by the controlled party is the same in all respects as steel manufactured by a comparable uncontrolled party. The controlled party would not be able to justify a materially different price for the same goods merely because its steel mill is much larger or much smaller. Admittedly, for a profitability comparison, the size of the enterprise could be one of the relevant factors but given the modular nature of software development units, a company which has a higher turnover might simply have more software units, and these would not necessarily bring significant economies of scale.

5.19 The onus was on the TPO/Assessing Officer to state and to show that the range of turnover sizes chosen by the assessee was wrong, and the TPO/Assessing Officer have neither given a finding to that effect, nor have they discharged that onus. Further, while conducting the bench marking process, the TPO has applied a turnover filter of Rs. 100 crores to Rs. 250 crores without providing sufficient reasons for applying such filters. Further, the TPO has not changed the upper end of the filter as applied by the assessee, but has not explained why he considered Rs. 100 crores as an appropriate lower end of the filter. This action of the TPO is not in the spirit of the provisions of law.

5.20 In this regard, the learned counsel further submitted that a TPO can reject the comparables in the TP Study of the taxpayer only if the comparables are insufficient or deficient. A mere variation in the turnover of the comparables does not alone qualify as a reason sufficient to reject functionally comparable companies. The above view is also supported by the decision of the Delhi Tribunal in Mentor Graphics (Nokia) (P.) Ltd.'s case (supra). At para 39.4 of the order, the Tribunal has observed as follows:

"The TPO could have carried fresh search only if the comparables drawn by the taxpayer was insufficient or had other deficiency."

5.21 As indicated earlier, the TPO has not given any reason as to why the comparables in the TP Study were insufficient/deficient and liable to be rejected. At this juncture, it would also be relevant to take note of the following observations of the Pune Bench of ITAT in the case of Egain Communication (P.) Ltd. (supra), at para 36, page 34 of the order :

"The learned CIT(A) was justified in taking entities having turnover between Rs. 5 crores to Rs. 25 crores but he was in error in considering turnover as the only relevant factor needed to be considered for a proper analysis. What about a large number of other factors which materially affect the profit ? The functions performed, assets employed, risk taken (FAR) analysis was also required to be undertaken as per the Transfer Pricing regulation and other guidelines. This was not done, which renders the comparison as unsound and unreliable."

Independent of the above, the TPO has considered M/s. Visualsoft Technologies Ltd. as a comparable company. As submitted during the course of the hearings, it should be noted that M/s. Visualsoft Technologies has significant research and development expenses (i.e., 4.61 per cent of the revenues) - refer to page 15 of the Annual Report for the period 2002-03, attached as Annexure 1. It should be noted that the assessee, being a captive service provider, does not incur any research and development expense, and moreover, is remunerated on a cost plus basis. Accordingly, the comparable, i.e. M/s. Visualsoft ought to have been screened out by the TPO as not being a comparable company. It should be noted that this test of 'research and development expense' has also been approved by the Kolkata Bench of the Tribunal in the case of Development Consultants (P.) Ltd. (supra) wherein the Tribunal has approved the rejection of companies having research and development expenses of more than 3 per cent of sales (refer para 31 of the said order.)

5.22 With regard to the above, the assessee would like to further place reliance on the decision of the Delhi Bench of the Tribunal in the case of Mentor Graphics (Nokia) (P.) Ltd. (supra), wherein it was held that depending on the facts of the case, the final set of comparables may need to eliminate differences by making adjustments; including an adjustment for R&D expenses. Accordingly, for a proper comparability, the TPO ought to have excluded M/s. Visualsoft as a comparable company.

5.23 The learned counsel had submitted for comparability under rule 10B, the focus is on the functions performed, and reference to other economic criterion are only in the context of the functions. This is clear from the provisions of rule 10B(2), the relevant extracts of which are reproduced below :

"(2) For the purposes of sub-rule (1), the comparability of an international transaction with an uncontrolled transaction shall be judged with reference to the following namely—

(b )the functions performed, taking into account the assets employed or to be employed and the risks assumed, by the respective parties to the transactions;

A mere reference to assets employed or the turnover, disregarding the functions performed would be an erroneous approach. In this connection, the learned counsel also drew reference to the decision of the Pune Bench of the Tribunal in the case of E-gain Communication (P.) Ltd. (supra). At para 36, page 34, of the order, the Tribunal has observed as follows :

"The learned CIT(A) was justified in taking entities having turnover between Rs. 5 crores to Rs. 25 crores but he was in error in considering turnover as the only relevant factor needed to be considered for a proper analysis. What about a large number of other factors which materially affect the profit? The functions performed, assets employed, risk taken (FAR) analysis was also required to be undertaken as per the Transfer Pricing regulation and other guidelines. This was not done, which renders the comparison as unsound and unreliable."

Further, the example of an automated vis-a-vis manual manufacturing unit as given by the DR is not supported by logic. The products from a handloom would be commercially different from those produced through a power loom. For example, a hand made sari would typically cost more as compared to a machine weaved sari. Also, in the case of the assessee, even if one were to compare the return on assets (i.e., capital employed), the assessee would very clearly pass the arm's length test. The ratio of PBIT over capital employed in the case of the assessee is 24.99 per cent, whereas the average in the case of the comparables approved by the CIT(A) works out at 26.66 per cent. It should also be noted that this argument of the DR of "return on assets" is being brought up for discussion for the first time.

5.24 The ld. DR has also drawn reference to the decision of the Delhi Bench of the Tribunal in the case of Ranbaxy Laboratories Ltd. (supra). At the first instance, it should be noted that the paragraph quoted by the DR is not an observation of the Tribunal in the case of Ranbaxy Laboratories Ltd. (supra), but a reproduction of the observation made by the Tribunal in the case of Mentor Graphics(Nokia) (P.) Ltd. (supra) at page No. 27, wherein the Tribunal has discussed various aspects relating to comparability analysis. Even in the said decision, the comment was made by the Tribunal in the context of the functions performed. This assertion of the DR contradicts the argument made by him that there was no search conducted by the TPO. The learned counsel submitted that the DR has not given any reasons for rejecting the comparables in the TP Study. Further there is no such observation in the show-cause notice issued by the TPO (refer pages 231 to 235 of the paper book). In fact, the show-cause notice and the order of the TPO are totally in different directions and are clearly independent of one another. In the show-cause notice, the TPO had proposed to compute the arm's length price by adopting a net profit margin of 16.82 per cent (based on an updated net margin working of the comparables on the TP Study). However in the order, the TPO proceeded to pass the order based on fresh comparables as selected by him, and arriving at a net margin of 21.14 per cent. Neither the TPO nor the Assessing Officer have at any stage of the proceedings established that any of the four conditions of section 92C(3) have been satisfied. Further the DR, at various occasions, has indicated that the assessee had 'selected' the comparables. The learned counsel reiterated that unlike the tax department, it has not 'selected' comparables.

5.25 The assessee had identified the appropriate category in the database of companies and had used filters for eliminating companies that are not comparable to it. The filters used for eliminating companies are system based (quantitative filters) and manual eliminations (qualitative filters). The final comparables are those companies that have survived all the elimination filters. On the other hand, the tax department has resorted to cherry picking comparables (typically those companies which are making huge profit margins) without explaining the basis and rationale thereof. This is also clear from the way the department has 'normalised' the profits of super-profit making companies.

5.26 The learned counsel submitted that the turnover range selected by it is not 'arbitrary'. Further, the Departmental Representative has not defined what is 'arbitrary' for him. A turnover filter (system based filter) of 100-250 crores (that the DR claims to be relevant in the instant case) if used now will give an initial set of around 20 companies only. At the time of doing the TP Study, the companies in the database under the heading 'software industry' gave an initial set of 402 companies. Various filters were applied on these 402 companies to arrive at the final set of comparables, one among them being the turnover filter of 5 to 250 crores. In this regard, it may be noted that Nasscom classifies companies Tier-wise, based on the sales revenue of the companies. For this purpose, Nasscom classifies companies as Tier 1 if the turnover of the company is below USD 100 million (approximately Rs. 1 to Rs. 400 crores). Further, Capitaline Plus database classifies software companies having a turnover in the range of Rs. 1 to Rs. 1,000 crores as 'small and medium size' software companies. Transfer pricing is not an exact science. This has also been noted by the Tribunal in the case of Mentor Graphics (Nokia) (P.) Ltd. (supra), where at para 24 of the order, the Tribunal has observed as follows :

"24. It is true that 'transfer pricing' is not an exact science, evaluation of transactions through which the process of determination is carried in an art where mathematical certainty is indeed not possible and some approximation cannot be ruled out, yet it has to be shown that analysis carried out was 'judicial' and was done after taking into account all the relevant facts and circumstances of the case."

Further, the OECD, with its vast knowledge and experience in the area of transfer pricing has also observed that :

"It should also be recalled at this point that transfer pricing is not an exact science but does require the exercise of judgment on the part of both the tax administration and taxpayer."

5.27 Considering the above background that transfer pricing is not an exact science and the fact that the assessee's turnover during the relevant year was Rs. 153 crores, it is grossly erroneous on the part of the DR to argue, that the turnover filter applied by the company was arbitrary . Moreover, the TPO ought not to have brought in a new turnover filter of Rs. 100 to Rs. 250 crores, especially considering that there was no infirmity in the filter applied by the assessee. The need for a larger sample of comparables has been established by the Tribunal in the case of Aztec Software & Technology Services Ltd. (supra ). One may draw reference to para 155 of the decision where the Tribunal has commented as follows :

"The sample size of data taken by the taxpayer, to support the ALP, was too small to come to any general conclusion."

It is quite evident that the sample size data selected by the TPO (i.e. 20 companies) was very small and not sufficient to apply any meaningful elimination filters for arriving at the final set of comparables. Even if one were to now do a broad search for a universal set of companies for 'computer software/software services and consultancy services' the average net margin would be in the region of 6.34 per cent (293 companies from PROWESS database as updated on 15-2-2008). The turnover filter applied by the TPO is without any basis. Further, the DR's computation argument of a range of ±60 per cent from the turnover of the assessee is not factually correct. At the lower end, the TPO has reduced 33 per cent and at the higher end, the TPO has increased 66 per cent for arriving at his filters. This itself shows the arbitrariness in the procedure of the TPO. Without prejudice to the submission that there should be no variation in the turnover filter applied by the assessee, even based on the upper turnover filter of 66 per cent applied by the TPO, if the same percentage is applied on the lower turnover filter, the turnover range would work out to Rs. 60 crores to Rs. 240 crores. The TPO had not disclosed his search process to the assessee. Having asserted that the TPO had not disclosed his search process to the assessee, the DR ought to have pointed out, where the disclosure was made. The reference made by the DR to the decision of E-gain Communication (P.) Ltd. (supra) to support the turnover filter of the TPO is incomplete and erroneous. The DR has deliberately quoted the decision selectively. In the said decision, the Tribunal has immediately thereafter also mentioned the following :

"But we see no justification for considering oversized companies as taken by the TPO. The teamed CIT(A) was justified in taking entities having turnover between Rs. 5 to Rs. 25 crores but he was in error in considering turnover as the only relevant factor needed to be considered for a proper analysis. What about a large number of other factors which materially affect the profit? The functions performed, assets employed, risk taken (FAR) analysis was also required to be undertaken as per the Transfer Pricing regulation and other guidelines. This was not done, which renders the comparison as unsound and unreliable."

The DR's argument on human assets are without any basis and are full of contradictions. One such example of contradiction is at page 18 of the written submissions of the DR, which has been extracted below :

"Moreover, the turnover in general is directly proportional to the number of employees as the billing in the software industry in general is based on man hourly basis. Therefore, turnover gives a close approximation to the number of employees as the data regarding the number of employees is not available in respect of all comparable companies selected by the TPO threw up closer comparables and hence, should be accepted as correct."

In the above comment, at one hand, the DR has mentioned that the turnover companies is directly linked to the number of employees and at the other hand, he himself has observed that the data relating to the employees is not available. Separately, it should be noted that the Indian accounting principles do not consider human resources as an asset to be quantified and included in the balance sheet. There is no basis for concluding that employee cost being in the same range of turnover would mean that asset bases are comparable. The DR can also consider the derived man hour billing rate for comparison purpose. The derived man hour billing rate of these assessee works out USD 29.24, as computed below :

The TPO in the assessment made for assessment year 2002-03 has herself used an industry benchmark of USD 18-25 per hour. If the industry rates are considered as a potential CUP, the man-hour rate of the assessee and the consequent value of the international transactions of the assessee with its associated enterprises would be at arm's length. The DR has completely missed the context while comparing the R&D Work of the assessee and that of M/s. Visualsoft. The assessee is an R&D Service provider and has provided these services to its clients. The R&D work carried out by the assessee is; a service to its associated enterprise for which the assessee is remunerated on a cost plus basis. The assessee does not have any R&D expenditure on the other hand, M/s. Visualsoft has conducted R&D for itself, and the R&D outflow is an R&D expense for M/s. Visualsoft. In the above context, it would be relevant to note that in the case of Mentor Graphics (Nokia) (P.) Ltd. (supra), the Tribunal has held the need for eliminating differences on account of R&D expenditure. The observations of the Tribunal has been reproduced below :

"Depending on facts of the case, final set of comparables may need to eliminate differences by making adjustments for the following :

(a) working capital.
(b) adjustment for risk and growth.
(c) adjustment of R&D expenses."

5.28 The learned counsel submitted that the request for re-computing the margins of M/s. Silverline cannot be accepted, inter alia, for the following reasons :

(i) The DR cannot request for increasing the margin and improving his position in an appeal which has been filed by the assessee. In fact, it would be relevant to note that even the CIT(A) has considered the margin of M/s. Silverline at (-) 11.37 per cent. It should be noted that no cross objection by the revenue on this matter.
(ii) Without prejudice to the above, the financials attached by the OR to the written submissions are not for the year ended 31-3-2003, and relate to 15 month period ended 30-6-2002.
(iii) Any such computation would not be in conformity with rule 10B(4) and rule 10D(4).
(iv) There are no provisions by which the margins can now be recomputed by the TPO.

While conducting the comparability analysis, the TPO rejected companies having substantial related party transactions. This view of the TPO was further approved by the CIT(A). The CIT(A) in her order has also commented that the criteria to be used to exclude companies having related party transactions should be 25 per cent on sale. In other words, the CIT(A) has confirmed the rejection of comparable companies which have related party transactions of more than 25 per cent over its sales. For the purpose of reaching her conclusion, the CIT(A) has referred to the transfer pricing assessments made in other cases. In connection with the above, the assessee draws attention to rule 10A(a) which defines the term 'uncontrolled transaction' to mean :

"a transaction between enterprises other than associated enterprises, whether resident or non-resident."

As is clear from the above definition, for the purpose of the comparability analysis, the comparables should not behaving transactions with its associated enterprises. In other words, a company having any related party transaction (i.e. even a single rupee of related party transaction) should not be considered as a comparable company, this view is also supported by the decision of the Delhi Tribunal in the case of Mentor Graphics (Nokia) (P.) Ltd. (supra). At para 35 of the said order, the Tribunal has observed that :

"In the first place, while screening and filtering of comparables, he took into account companies which were dealing with their related companies either as subsidiaries or as parent companies. Thus, controlled transactions were taken into account which is against the very basics of the transfer pricing guidelines."

The above view is also supported by the OECD. In the draft notes on comparability issued by the OECD, it has been observed as follows :

"The working party discussed these issues and concluded that controlled transactions should by no means be used as the basis for a transfer pricing adjustment. . . ."

Considering the above, the learned counsel had submitted that the companies which have been a single rupee transaction with associated enterprises should not be considered as a comparable company. This issue has also been acknowledged by the CIT(A) and at page 48 of her order, she has observed as below :

"With due respect of the Hon'ble ITAT, I would also submit that while in a perfect world, comparables should be taken that have no related party transactions, in practice, finding such comparables appears to be bordering on the realm of impossibility. I am of the view that the approach to be taken should be one where comparables should be found where the extent of the related party transaction is not of such an order of magnitude as to distort the results and therefore made comparison less than meaningful. Such I believe is the approach that I have taken and the approach the assessing authorities have taken as well."

In view of the above submissions and based on the provisions of the law, the learned counsel had argued that the TPO and the CIT(A) have erred in including companies having any related party transactions. Without prejudice to the above contention, even if companies having related party transactions were to be included in the set of the final comparables, the filter of 25 per cent over sales is ad hoc, and without any basis. Further, even if it were be assumed that companies with related party transactions could be included on the basis of the 25 per cent filter, the learned counsel submitted that the TPO/CIT(A) should have then also rejected two more companies, namely Hinduja TMT Limited [margin computed by the CIT(A) at 111.87 per cent, normalized to 32.71 per cent] and Xansa India Limited [margin computed by the CIT(A) at 25.78 per cent] on the basis that these companies have related party transactions to the tune of 27 per cent and 99 per cent of their sales, respectively. Accordingly, in the event the ITAT is of the view that :

(a) companies with related party transactions can be considered as comparables; and
(b) 25 per cent of sales is an appropriate filter for this purpose,

the learned counsel had submitted that the above noted two companies should also then be excluded from the set of comparable companies. As submitted above, the above matter is clearly covered by the provisions of rule 10A(a), and the decision of the Tribunal in the case of Mentor Graphics (Nokia) (P.) Ltd. (supra), both of which establish that for the purpose of the comparability analysis, the comparables should not be having any transactions with its associated enterprises. At one end, the department has thought it fit to relax the conditions of rule 10A relating to controlled transactions, for getting a larger number of comparables. On the other hand, the TPO has altered the ranges of turnover filter applied by the taxpayer, which has resulted in a reduction of the sample size of comparables. In other words, when it suits the DR/TPO, they seek shelter under practical reasons for tweaking the law to suit their convenience. Further, the reference drawn by the DR to various other sections [such as section 40A(2)(b) and section 92A(2)(a)] to justify the selection of 25 per cent as the cut off is misplaced. In fact, as submitted above, the provisions of rule 10A(a) clearly prove that where the Legislature intended to give a cut off they had provided the same in the relevant section itself. Had the Legislature intended to provide any such cut-off for the purpose of eliminating companies with rented party transactions, they would have provided the same in the statute itself. Instead, the Legislature has clearly provided under rule 10A that companies with related party transactions cannot be considered as 'comparables'. As submitted earlier, the assessee had followed a methodical search process in the TP Study. Had the data been available at the time of conducting the TP Study, the assessee concedes that the two companies would have got rejected. This is not the appropriate stage to remove these two companies.

5.29 Rule 10B(1)(c)(iii )/(e)( iii) provides that the margins of the comparable companies need to be adjusted to take into account the functional and other differences which could materially effect such margins in the open market. During the course of the proceedings, the assessee had requested the CIT(A) to grant an adjustment to the margin of the comparable companies on account of the difference in the risk profile of the comparable companies. However, the CIT(A) has not approved the claim of the assessee on the basis that:

(a) the assessee is bearing a significant risk in the form of a single customer risk; and
(b) the method for claiming the adjustment of 5.25 per cent is ad hoc and without basis.

In connection with the above issue, the learned counsel submitted that its claim is squarely within the provisions of the law (i.e. rule 10B), and the CIT(A) has not given any legal basis for the rejection of the claim. Further, the CIT(A) has grossly erred in reaching a conclusion that the assessee bears a significant risk in the form of a single customer risk. The CIT(A) ought to have appreciated that the asssessee is working on a cost plus basis, and accordingly, is insulated from all kinds of business and operational risks. The claim of the assessee for an adjustment on account of differences in risk profile is also supported by the decision of the Delhi Tribunal in the case of Mentor Graphics (Nokia) (P.) Ltd. (supra). The observation of the Tribunal in paragraphs 27, 27.1 and 33 of the order is reproduced below :

"27. After the selection of the comparables, best method of determining Arm's Length Price is selected. Thereafter, functional analysis is carried to identify functions, risk and assets of uncontrolled transactions and comparison is carried with characteristics of the controlled transaction. This is necessary to find whether comparable selected are really comparable and reliable. Comparison based on functional analysis include economically significant activities and responsibilities undertaken or to be undertaken by the independent and associated enterprises. The structure and organisation of the group and more particularly the judicial relationship between different entities of same group are to be seen. The function that need to be identified while carrying comparison as per OECD guidelines include design, manufacturing, assembling, research and development, servicing, purchasing, distribution, marketing, advertising, transportation, functional and management activities. It is also necessary to examine as to what is the principal function of the entities. The analysis of comparison should consider total assets employed and assets used to earn profit. The risk assumed by respective parties is a very important consideration. It in a simple principle of economics that the greater the risk, the greater the expected return (compensation). If there are material and significant differences in the risk involved then the comparable identified are not correct as appropriated adjustments for differences in such cases are not possible. Therefore, while performing searches for potential comparable companies, not only turnover and operating profit but functions performed and risk profile are also to be considered. However, it can always be shown on the given facts of the case that comparable found are similar or almost similar to the controlled transaction and no adjustments are needed. It is useful to see the level of intangible assets in comparable to an appropriate base. Depending on facts of the case, final set of comparables may need to eliminate differences by making adjustments for the following :

(a) working capital
(b) adjustment for risk and growth
(c) adjustment of R&D expenses.

27.1 The risk not only due to human resources, infrastructure and quality which are normally taken into account yet more significant risks like market risk, contract risk, credit and collection risk and risk of infringement of intellectual property are being ignored here.

33. Most of the business risk such as contract risk, market risk, credit risk, warranty risk, price risk etc. were essentially borne by parent AE."

This matter has also been considered by the Pune Bench of the Tribunal in the case of E-gain Communication (P.) Ltd. (supra). Attention is drawn to the comments of the Tribunal at para 40 of the order, which is reproduced below :

"40. We further agree with the contention of the learned counsel for the tax-payer that the benefit of adjustment was required to be given in working the margin of profit of the taxpayer for not undertaking any risk in the transactions involved with its parent company. However, evaluation of above risk in the present case is not necessary as even otherwise the margin of profit shown by the taxpayer has fully satisfied the Arm's Length Price benchmark."

The contention of the DR that the assessee is bearing any single customer risk is factually incorrect. During the period 2002-03, the assessee had entered into international transactions with 57 associated enterprises (see pages 323 to 330 of the paper book.)

5.30 The DR has outweighed the 'single customer risk' against the other risks which are borne by the comparable companies (which are not captive service providers). As discussed above, the Tribunal in cases of Mentor Graphics (Nokia) (P.) Ltd. (supra) and E-gain Communication (P.) Ltd. (supra) have clearly established the need for making an adjustment for difference in risk profile. The DR also ignored the following :

(i) the assessee is working in a cost plus model and is earning constant/consistent profit in the current and past years, and is also likely to make consistent profit in the foreseeable future.
(ii) the assessee's quantum of business has increased considerably. The revenues from software development services in the financial year 2001-02 was Rs. 137 crores which increased to Rs. 207 crores in the financial year 2003-04.
(iii) the levels of risk would be significantly different as between service provider who has a single third party customer and another service provider who has a single (or for that matter, even multiple) captive customers.

To explain the above by way of an example, if two manufacturing companies are considered (i) ABC Ltd. which is a captive contract manufacturer supplying corrugated boxes only to its parent company in USA, namely ABC Inc. and (ii) XYZ Ltd. which is an independent company supplying identical corrugated boxes only to ABC Inc. (i.e. ABC Inc. is an independent client for XYZ Ltd.). Apart from this difference, both ABC Ltd. and XYZ Ltd. operate identically with respect to the following :

(a) Both supply their products only to one customer.
(b) Both have a fully automated manufacturing facility.
(c) Both source their raw materials from the same vendor.
(d) Both have the same number of workers with the same skill sets.
(e) Both have the same volume ( i.e. sales) during a year.
(f) Both enjoy the same tax benefits.

After considering all aspects of the business of ABC Ltd. and XYZ Ltd., it would be anybody's guess that the risk of ABC Ltd. would be less than that of XYZ Ltd. It can be no one's case that when a person would render services to a single customer, he would charge a premium for his services. In fact, the assertion made by the DR is not in line with; any theory of economics. This fact is also emphasized in the written submissions made by the DR in the context of 'single customer' risk. The assessee bears very nominal/nil business risks. The DR has drawn reference to the commercial agreement at page 152 of the paper book. Clause 5 of the very same agreement clearly provides that all risks are of the customer. The relevant extract has been reproduced below :

"5.1 The ownership and copyright of all deliverables will be held by Philips Industrial Activities, N.V., at the end of the assignment. All risks are to the account of Philips Industrial Activities, N.V."

In the first instance, it should be noted that there has been a consistent growth in the quantum of business of the assessee. The revenues from software development services in the financial year 2001-02 was Rs. 137 crores which increased to Rs. 207 crores in the financial year 2003-04. Since the assessee is operating on a cost plus model, the margin remains unchanged irrespective of the quantum of business. Independently, while making a comment that may impact on the associated enterprise would have a serious impact on the assessee, the DR has failed to appreciate that the growth of any business is linked to the economy as a whole. The DR's argument that "the captive service provider is also affected by the market condition indirectly" is simply seeking to blur the issue at hand, any decline/slow down in the economy will affect business, be it with a single customer or multiple customers. If the growth of the associated enterprise slows down, no doubt, the flow of business to the captive service provider could also come down. However, the associated enterprise would still continue to compensate the captive service provider for such reduced work at full cost plus a reasonable mark up.

5.31 Compare this situation with the associated enterprise giving work to a third party service provider during a slow down. The associated enterprise might bargain hard on the price and the third party service provider might even agree a lower price just to keep his business going during the slow down. The fact that a captive; service provider (a separate and distinct legal entity in India) is assured of business from an AE (another legal entity outside India) and is compensated at a consistent reasonable mark up over costs incurred by it (i.e. the captive service provider) is sufficient to demonstrate that the captive service provider is effectively insulated from all business risks and, as such, any mark up ought to be only commensurate with the captive service provider's risk taking ability, and not beyond. During the course of the hearing, the DR referred to the commercial agreements (page 153 of the paper book) to demonstrate that the assessee also carries on pricing risk as there is an overall cap on the fee the assessee can charge to the assessee.

5.32 The learned counsel on the argument of the ld. DR pointed out is incorrect as it is based on an ignorant interpretation of the said agreements. This can be clearly established as below :

(a) Clause 1( a) of the agreement at page 153 of the paper book very clearly provides that the invoice value will be arrived at by adding a 10 per cent margin to the total cost for a particular month.
(b) Clause 2( a) of the agreement at page 153 of the paper book only documents that on a realistic estimate, the fee to be charged for the assignment would be Euro 25,000. The said clause nowhere restricts the fee that can be charged by the assessee under clause 1( a) of the said agreement. The purpose of clause 2( a) of the agreement is only to provide a framework to both the parties for the assignments.
(c) In any case, the agreement represent only two of the various agreements entered into by the assessee with its associated enterprises of assessment year 2003-04.
(d) The DR can also consider the derived man hour billing rate for comparison purpose. The derived man hour billing rate of the assessee works out USD 29.24.

Further, the TPO in the assessment made for assessment year 2002-03 has herself used an industry benchmark of USD 18-25 per hour. If the industry rates are considered as a potential comparable uncontrolled price (CUP), the man hour rate of the assessee and the consequent value of the international transactions of the assessee with its associated enterprises would be at arm's length.

5.33 The learned DR's act of giving a new list of 25 supposedly captive companies is an example of making irresponsible and incorrect assertions. For instance, the DR has quoted, M/s. Wipro as a captive, which is wholly erroneous. In this regard, the learned counsel would like to draw reference to the following paragraphs from the draft OECD comparability notes :

Page 24 : "9. Many business commentators have expressed strong concerns with regard to the sue of "secret comparables", which they regard as unfair and lacking transparency. They urge the OECD to specifically reject the use of so-called "secret comparable" as contrary to the obligation on both taxpayers and tax administrations to make a "good faith showing" that the determination of transfer pricing is consistent with the arm's length principle."

Page 26 : "18. In addition, if the selection process is not transparent to the taxpayer, it cannot verify whether or not the process is subjective/objective, reliable/unreliable, therefore whether the result is right or not. Because the identity of the comparable is not disclosed to the taxpayer, the taxpayer cannot evaluate the objectivity/reliability of the selection process from the result."

Based on the above, the learned counsel submitted that the data provided by the OR in his submissions is; clearly 'secret comparables' and cannot be considered during the course of these proceedings. The final comparables that were accepted by the Tribunal in the Mentor Graphics (Nokia) (P.) Ltd.'s case (supra) and E-gain Communication (P.) Ltd.'s case (supra) clearly demonstrate that there are many other companies that are making net profit margins in the range of 5 to 10 per cent. In fact, the margins of the assessees in both these cases were in the range of 5 per cent to 7 per cent. The list of 25 companies is new material produced for the first time before the Tribunal. No such argument was raised during the course of assessment proceedings. The DR has himself confessed that the data is available only in private domain. Further, though the DR has argued that no fresh search has been conducted by the TPO, the action of producing a fresh list of 25 companies is contrary to such an argument. The companies listed include entrepreneurial companies such as Wipro, Emphasis, Subex. Further, this argument suffers from various infirmities such as the data submitted is for financial year 2004-05 (whereas the current appeal is for financial year 2002-03).

5.34 The learned counsel submitted that the assessee understood that there are a significant number of other captive companies who have earned margins in the region of 5 to 10 per cent. The DR has purposefully not highlighted such companies, which clearly brings out the anomalies in the process of cherry choosing. Even if one were to now do a broad search for a universal set of companies for 'computer software services and consultancy services' the average net margin would be in the region of 6.34 per cent (293 companies from Prowess database as updated on 15-2-2008). Margin of entrepreneurial companies cannot be compared with the assessee, who is a captive service provider bearing minimal/no risks, without any adjustments. The approach of the DR seems to be bias as there are many other companies whose margins are in the range of 5 to 10 per cent. The DR has ignored such companies while making his argument. The margins of these companies cannot be directly taken as a benchmark without doing a proper FAR analysis to eliminate differences on account of functions, assets and risks. The DR has himself accepted that the companies listed are earning margins better than the industry average. This clearly shows that the DR is highlighting only the companies at the higher end of the spectrum, and that too, without doing a proper FAR analysis.

5.35 The DR has argued that the assessee bears market risk as it has incurred Rs. 62.77 lakhs of advertisement and publicity expenditure during the year. The learned counsel places on record that the expenditure of Rs. 62.77 lakhs essentially include donation, sponsorship and other expenditure relating to educational contributions and advertisement for recruitment. The above expenditure has been incurred to create awareness with a view to attract potential employees and has nothing to do with market risks. Independent of the above, since all of the services of the assessee are rendered to its associated enterprise, there is no marketing risk borne by the assessee. This has also been established in the TP Study of the assessee (refer page 74 of the paper book). It should also be noted that the assessee does not have any marketing department. In any case, the assessee operates on a cost plus model and all costs are included in the cost base for recovering from the service recipients.

5.36 The learned counsel submitted that the 'risk return trade off' is not a hypothesis, but a simple and well accepted principle of economics. The Tribunal in case of Mentor Graphics (Nokia) (P.) Ltd. (supra) has also confirmed this proposition. Relevant extracts of the said decision are reproduced as under :

"27. ... It is a simple principle of economics that the greater the risk, the greater the expected return (compensation)."

With regard to objection raised by the DR relating to increase of mark up from 6 per cent to 10 per cent the learned counsel had submitted that the increase in the mark up of 5 per cent to 10 per cent was in view of the additional functions performed by the assessee (refer page 2:22 of the paper book). The learned counsel submitted that its bona fide action during the financial year 2002-03 to synchronise its margin with the prevailing arm's length criterion under Indian market conditions cannot, by itself, become a basis for alleging that "the taxpayer had not fixed the price of international transactions at arm's length, but the same were fixed to suit the groups-convenience." It should also be noted that the margins of the assessees in the cases of Mentor Graphics (Nokia) (P.) Ltd. (supra) and E-gain Communication (P.) Ltd. (supra) were also in the range of 5 per cent to 7 per cent, which was approved as being at arm's length by the Tribunals.

5.37 The reference of the DR to various studies is out of context. The ld. DR had made a sweeping statement which is based on his conjecture and surmise. MNCs recognized the fact that international operations are subject to various risks, such as political, economic, different legal systems, etc. which could also adversely affect their operations and increase their overall risk of doing business. The DR seems to have incorrectly assumed that by setting shops across the globe, the MNC has transferred some of its risk to the new legal entity in the foreign jurisdiction. The DR has failed to prove as to how by multi-nationalising their own operations, the reduction of external/market risks to the MNC Groups on an overall basis would actually results in increasing the risk of the captive service provider in the foreign jurisdiction. The DR's comments such as :

(i) the profit margin of a company is not dependent on the marketing expenditure etc.
(ii) the concept of arithmetical mean takes care of differences arising in the margins of the comparables on account of various factors including risk profile; and
(iii) the credit risk is; neutralized on account of the bad debt debited to the profit and loss account of the comparable concerned.

are devoid of logic and without any commercial or legal basis.

5.38 The Delhi Bench of the Tribunal in the case of Rolls Royce Plc. v. Dy. DIT [2008] 19 SOT 42 has indicated that marketing is a primary activity for earning profit, and but for such marketing being vigorously undertaken, the profit would not accrue. Further, the DR has not appreciated the purpose of arithmetical mean [proviso to section 92C(2)] and the adjustments to be made to the margins rule 10B. The proviso to section 92C(2) provides an option to the taxpayer to consider the arithmetic mean of the prices determined. On the other hand, rule 10B(1)(c)( iii)/(e)( iii) provides that the margins of the comparable companies needs to be adjusted to take into account the functional and other differences which could materially effect such margins in the open market. Accordingly, in addition to the adjustment for arithmetic mean, adjustments on account of risk are clearly envisaged in the Rules, and both co-exist.

5.39 Out of the final set of comparable companies [as approved by the CIT(A)], the CIT(A) has identified two companies as being companies earning super profits. The CIT(A) has 'normalised' the margins of such companies, by replacing their profit margins with the profit margins of the next highest profit making company from the set of final comparables.

5.40 Rule 10B(1)(c)(iii )/10B(1)(e)( iii) allows an adjustment to the margins of the comparable companies to take into account differences which could materially affect the amount of profit margins in the open market. The CIT(A) has acknowledged that the two companies have made super profits. As a corollary, it implies that these two companies are enjoying certain extraordinary benefits/advantage over the other companies, as a result of which these companies are earning significantly higher profit margins which no other company in the given industry earns, under ordinary circumstances.

5.41 Rule 10B(3) provides that an uncontrolled transaction shall be considered as a comparable if :

(a) none of the differences between the comparable company and the controlled transaction are likely to materially affect the profit arising from such transactions in the open market; or
(b) reasonably accurate adjustments can be made to eliminate the material effect of such differences.

In the instant case, it was acknowledged by the CIT(A) that two companies in the set of comparables as approved by the CIT(A) are making super profits. The super profits of these companies would materially affect the comparability analysis. The learned counsel had submitted that in the absence of any specific provision for making the necessary/appropriate adjustment to the margins of the super profit making companies and the ability to accurately identify and quantify the adjustments to eliminate the differences, the two profit making companies should not be considered as 'comparable'. Therefore, given the fact that a company is making super profits, such a company cannot be taken as a comparable company. There is no provision allowing the consideration of such super profit making company as a comparable company. Further if at all, such a company were to be considered as a comparable, such companies would need appropriate adjustments to account for the material difference as provided in rule 10B(1)(c)(iii )/10B(1)(e)( iii). The rule does not envisage any normalization of the margins of such super profit making companies on an ad hoc basis.

5.42 The CIT(A) has normalized the super profit margins of such companies using the next highest profit earned by a company in the set of comparables. This approach followed by the CIT(A) to normalize the super profit margins is clearly biased as this method would result in increasing the average or the arithmetical mean margin of the comparable companies. Without prejudice to the above, at one extreme, the assessee can also counter argue that the CIT(A) ought to have normalized the margins of these two companies using the margin of the lowest profit making company in the set of comparables as approved by the CIT(A).

5.43 The issue of exclusion of high profit making companies has been addressed by the Tribunal in the cases of Mentor Graphics (Nokia) (P.) Ltd. (supra) as well as E-gain Communication (P.) Ltd. (supra).

In the case of Mentor Graphics (Nokia) (P.) Ltd. (supra), the Tribunal has observed as follows :

"46.1 We are not taking into account high profit or high loss making companies as comparables."

Further, in the case of E-gain Communication (P.) Ltd. (supra), the Tribunal has, as para 37 of the order, commented as below :

"A cursory look at the chart in the assessment order of 20 comparables would reveal that the margin of profit shown by Thirdware Solutions Ltd. and WTI Advanced Technology is extraordinary at 67.65 per cent and 54.72 per cent respectively."

5.44 The DR has mentioned that the TPO has taken a liberal view by way of normalizing the margins of the companies accepted to be making super profits. The claim made by the DR is totally wrong. Despite there being no statutory provisions for normalization, the TPO has normalized the profits of these two companies; using the next highest profit earned by a company in the set of comparables. This approach followed by the TPO to normalize the super profit margins is clearly biased as this method has resulted in increasing the average or the arithmetical mean margin of the comparable companies. This method, in no way, can be said to be liberal. The assessee cannot be dependent upon a charitable whim of the TPO. The comparability analysis should be made on the basis of the provisions of law and not on an ad hoc basis. The argument of the DR that the act of normalization off setting for the need for risk adjustment is devoid of any merits and logic. The DR's comment of normalizing using the margin of Visualsoft as 28.61 per cent is erroneous. In fact, the normalization has been done by using a margin of 32.71 per cent. It would be out of place to mention here that in the pretext of normalization of super profits, the TPO has actually increased the margin of Infotech from 31.78 per cent to 32.71 per cent. There cannot be a more glaring example of a deliberate action by the TPO to arrive at disproportionately higher arithmetical mean with the objective of making a transfer pricing adjustment. Independent of the above, by a parity of reasoning, even M/s. Visualsoft, which has made a margin of 32.71 per cent should be considered as a company making super profits and accordingly not to be included.

5.45 The assessee follows a depreciation policy, whereby, the rates at which it charges depreciation on its assets is much higher than the rates at which the comparable companies charge depreciation. Accordingly, due to the difference in the rates of depreciation, the profit margins of the comparable companies are much higher than that of the assessee. Accordingly, in order to conduct a fair and equitable comparison of the profit margins in terms with rule 10B(1)(c)( iii) and 10B(1)(e)( iii), in the TP Study, the assessee made adjustments to the depreciation rates charged by the comparable companies. The CIT(A) has observed that in order for a comparison to be meaningful, adjustments on account of differences in accounting practices are necessary and are acceptable in principle. How- ever, she has rejected the claim of the assessee for an adjustment for the difference depreciation policy followed on the basis that the method adopted by the assessee for making the adjustment was too 'simplistic'. In relation to the above, the learned counsel submitted that the CIT(A) is of the opinion that making an adjustment based on rates of depreciation is too simplistic, a more systematic basis of making the said adjustment needs to be done. The CIT(A) has not expressly rejected the need for making such adjustment in the instant case. In such a scenario, if the CIT(A) believed that the methodology of the assessee was too simplistic, the CIT(A) ought to have computed the adjustment using a method which she found appropriate. At this point, the assessee would also like to highlight that the TPO had normalized the margins of super profit making companies by simply using the profit margin of the next highest profit making company. The approach of the TPO was also approved by the CIT(A). Neither the Act, nor the Rules, prescribe any method for normalizing super profit margins of companies. At one end, the CIT(A) has termed the depreciation adjustment made by the assessee to be too simplistic and at the other end, the CIT(A) has herself approved the TPO's process of normalizing the margins of super profit making companies, which is very simplistic. Further, it would be relevant to note that the concept of adjustment on account of different depreciation polices being followed has been approved by the Tribunal in the case of E-gain Communication (P.) Ltd. (supra). The DR has acknowledged the need for making an adjustment, but has not agreed to the methodology adopted by the assessee.

5.46 During the appellate proceedings before the CIT(A), considering the financial position of the assessee and the comparable companies, a request was made to the CIT(A) to allow making an adjustment on account of the difference in the working capital position. The learned counsel for the assessee also submitted that a similar working capital adjustment was also granted by the TPO in the succeeding assessment year, in the assessee's own case. However, the CIT(A) did not allow an adjustment for working capital citing that the same cannot be granted merely because such an adjustment has been allowed by the TPO for the succeeding assessment year. The CIT(A) has not addressed the merits of the case of the assessee. It should be noted that the assessee is a captive contract service provider, rendering software development services to its associated enterprises. The payment cycle in the case of the assessee is shorter as against a considerable time lag in case of a third party service provider. In case of the comparables as approved by the CIT(A), the holding period for working capital is around 8 months as against that of the assessee of around 3 months. Accordingly, the learned counsel submitted that considering the provisions of rule 10B(1)(c)( iii)/(e)( iii), an adjustment on account of working capital difference is required to be made for an equitable comparison. The need for making such an adjustment has also been approved by the Delhi Bench of the Tribunal in the case of Mentor Graphics (Nokia) (P.) Ltd. (supra). The observation of the Tribunal in para 27 of the order has been reproduced below :

"The risk assumed by respective parties is a very important consideration. It is a simple principle of economics that the greater the risk, the greater the expected return (compensation). If there are material and significant differences in the risk involved, then the comparable identified are not correct as appropriated adjustments for differences in such cases are not possible. Therefore, while performing searches for potential comparable companies, not only turnover and operating profit but functions performed and risk profile are also to be considered. However, it can always be shown on the given facts of the case that comparable found are similar or almost similar to the controlled transaction and no adjustments are needed. It is useful to see the level of intangible assets in comparable to an appropriate base. Depending on facts of the case, final set of comparables may need to eliminate differences by making adjustments for the following :

(a) working capital.
(b) adjustment for risk and growth.
(c) adjustment of R&D expenses."

The OECD in 'a public invitation to comment on a series of draft issues notes', has clearly recommended a need for making a working capital adjustment if the working capital position of the taxpayer is favourable as compared to the comparable companies. A brief computation of the working capital adjustment, as applicable to the present case and based on the comparables as approved by the CIT(A), is shown below :

Computation of working capital adjustment

Step 1 : L Net working capital for financial year 2002-03 for each of the comparable and the assessee is computed as under :

(Net working capital for financial year 2002-03) × 360 days=

Holding period

Sales for financial year 2002-03

Step 2 : Net holding period is computed

Net holding period = (Average holding period of comparables) - (Holding period of the assessee.)

Step 3 : The net holding period is applied to the Prime Lending Rate (PLR) to compute to the working capital adjustment.

[Net holding period (in days)] × PLR

360 days.

Actual workings are as under:

Step 1 : Holding period for comparables [all comparables approved by the CIT(A)] = 265.98 days.

Holding period for the assessee = 80.42 days.

(Note: None of the comparable companies have a holding period which is less than that of the assessee)

Step 2 : Net holding period is (265.98 days - 80.42 days) = 185.56 days.

For computing the working capital adjustment, instead of using the PLR, the following other returns on capital can also be considered.

The working capital adjustment as computed above is on the lines of the draft guidance on comparability issued by the OECD.

5.47 The assessee is a captive contract service provider, rendering software development services to its associated enterprises. While rendering services to its associated enterprises, the assessee does not bear any business and operational risks. Though in the order, the CIT(A) has acknowledged that there is a need for making an adjustment on account of the difference in the risk profile of the assessee, she has not granted the adjustment on the ground that the claim is "ad hoc and wholly without any basis". In this regard, the learned counsel had submitted that an adjustment on account of the difference in the risk profile can be made as under:

(a) The adjustment which is sought to be made is on account of difference in the risk profile of a captive service provider (assuming less than normal risks) and full fledged entrepreneurs (bearing all business and operational risks;)- In this regard, the economic theory of 'higher the risk higher the returns' would apply.
(b) Thus, where a risk insulated captive service provider would earn lower returns, a full- fledged entrepreneur would earn higher returns. The difference between the two returns would be the risk premium or the additional return which one would earn by bearing higher risks.
(c) Drawing the above theory to the financial markets, a normal risk bearing rate: of return could be the rate as notified by the Reserve Bank of India (RBI) being the Prime Lending Rate (PLR). This rate is primarily the base rate at which all commercial banks in India would ideally lend money to independent customers. During the previous year relevant to assessment year 2003-04, the PLR as notified by the RBI was around 11.50%.
(d) In comparison to the PLR, the RBI also notifies a bank rate which can be considered as a risk free rate of return. During the previous year relevant to assessment year 2003-04, the bank rate as notified by the RBI was around 6.25 per cent.
(e) Based on the above, the PLR can be considered as a normal risk beating rate of return whereas the Bank rate can be considered as the risk free rate of return. The difference between the two rates would be the return one would earn by bearing additional risks. Thus, the difference in the return would be around 5.25 per cent ( i.e. 11.50-6.25 per cent)
(f) Accordingly, the arm's length price as determined by the CIT(A) (which is a risk bearing rate of return) would have to be adjusted by the risk premium of 5.25 per cent. As all alternative to the above, if one were to consider the rate at which banks charge credit card holders (35.40 per cent) as the risk bearing rate, the risk adjustment required would work out to 29.15 per cent.

5.48 At the first instance, the learned counsel submitted that the need for a risk adjustment has been approved by the Tribunals in the case of Mentor Graphics (Nokia) (P.) Ltd. (supra), and E-gain Communication (P.) Ltd.'s case (supra). Independent of the above, the DR ought to appreciate the conceptual difference between the PLR and the Bank Rate. While the Bank Rate is the rate at which the RBI lends money to various banks, the PLR represents a base rate for loans disbursed by banks to its customers. During the course of the hearing, the DR has himself pointed out that the banks typically give loans at a rate which is higher than the PLR, and this additional difference is based on the bank's assessment of the risk profile of the borrower.

5.49 Considering the above, the assessee has been fair in requesting for an adjustment of 5.25 per cent which is the difference between the PLR and the Bank Rate, as a risk premium that an entrepreneur could earn as compared to a captive service provider. The correct risk premium would be much higher, since many companies are earning super profits, way beyond a risk free return. However, to limit controversy, the assessee has not taken a rate higher than the PLR while calculating the risk adjustment.

5.50 The proviso to section 92C(2) of the Act provides that while arriving at the arm's length price, at the option of the assessee, a price which may vary from the arithmetical mean by 5 per cent, can be considered as the arm's length, price. The proviso to section 92C(2) is reproduced as under:

"Provided that where more than one price is determined by the most appropriate method, the arm's length price shall be taken to be the arithmetical mean of such prices, or, assessment the option of the assessee, a price which may vary from the arithmetical mean by an amount not exceeding five per cent of such arithmetical mean." In this regard, relevant extracts form the explanatory memorandum and notes to clauses of the Finance Bill, 2002, which explained the 5 per cent variation, are produced here under : Extract form Explanatory Memorandum to Finance Bill, 2002 "Under the existing provision contained in the proviso to sub-section (2) of section 92C of the Income-tax Act, if the application of the most appropriate method leads to determination of more than one price, the arithmetical mean of such prices shall be taken to be the arm's length price in relation to the international transaction. With a view to allow a degree of flexibility in adopting an arm's length price, it is proposed to amend the proviso to sub-section (2) of the said section to provide that where the most appropriate method results in more than one price, a price which differs from the arithmetical mean by an amount not exceeding five per cent of such mean may be taken to be the arm's length price, at the option of the assessee."

Extract from Notes on Clauses to Finance Bill, 2002.

"Under the existing provisions contained in sub-section (2) of the said section 92C, the most appropriate method shall be applied for determination of arm's length price in the manner prescribed. The proviso to sub-section (2) provides that if the application of the most appropriate method leads to determination of more than one price, the arithmetical mean of such prices shall be taken to be the arm's length price in relation to the international transaction, it is proposed to substitute the aforesaid proviso to sub-section (2) so as to provide that where the most appropriate method results in more than one price, the arithmetical mean of such prices or, at the option of the assessee, a price which differs from the arithmetical mean by an amount not exceeding five per cent of such mean may be taken to the arm's length price in relation to the international transaction." It was submitted that the language of proviso to section 92C(2) is clear and even without reference to extracts from 'Explanatory Memorandum to Finance Bill, 2002' and 'Notes on Clauses to Finance Bill, 2002', no different view can be formed. The only criterion for availing itself of the 5 per cent variation is that the same is available at the option of the assessee. Based on the above, one can reasonably conclude that a 5 per cent variance is allowed as a standard deduction which should be allowed while making an adjustment to the arm's length price, as declared by the assessee."

Vide Circular No. 12-2001, dated 23-8-2001, the CBDT had instructed its officers to not make any adjustment to the arm's length price as determined by the taxpayer, in case the same is up to 5 per cent (less or more) of the price as determined by the Assessing Officer. Relevant extracts of the said Circular are reproduced as under :—

"However, this is a new legislation. In the initial years of its implementation, there may be room for different interpretations leading to uncertainties with regard to determination of arm's length price of an international transaction. While it would be necessary to protect our tax base, there is a need to ensure that the taxpayers are not put to avoidable hardship in the implementation of these regulations. In this back ground the Board have decided the following—

(i )The Assessing Officer shall not make any adjustment to the arm's length price determined by the taxpayer, if such price is up to 5 per cent less or up to 5 per cent more than the price determined by the Assessing Officer. In such cases the price declared by the taxpayer may be accepted."

The following should be noted in the context of the above noted circular:

(a) The said circular does not address the treatment in a case where the arm's length price determined varies by more than 5 per cent; and
(b) The Circular was issued by the CBDT in August 2001 whereas the proviso to section 92C was amended vide Finance Act, 2002, to introduce the concept of 5 per cent deduction to the arm's length price.

Thus, the above noted circular is not relevant in the context of the issue raised. In summary—

The proviso to section 92C(2) provides that in the event that more than one arm's length price is determined, the arm's length price would be :

- the mean of such prices; or
- at the option of the assessee, a price which varies by 5 per cent of such mean.

The above clearly establishes that the only condition for substituting the mean by a price which varies by 5 per cent of such mean is that the same is at the option of the assessee. Accordingly, if any transfer pricing adjustment is to be made, if the assessee opts, the adjustment can be computed with reference to the 95 per cent of the mean price. At this juncture, it would be relevant to note that this view has in-principal been approved by the Kolkata Bench of the Tribunal in the case of Development Consultants (supra) (refer para 22 on page 16 of the said order.) In the instant case, the TPO/CIT(A) have not applied the proviso to section 92C(2) i.e. they have not allowed a standard deduction of 5 per cent from the arithmetical mean and have thus ignored a specific and mandatory provision of law.

5.51 The learned counsel had submitted that his arguments are squarely covered by the decision of the Tribunal in the case of Development Consultants (supra). In the said case, the Tribunal, after considering all the facts and issues of that case had approved the methodology of the assessee in that case. The discussion on this matter is covered under paragraphs 22 and 23 of the said order, which has been reproduced below :

"22. The learned AR submitted that in case of assessment year 2003-04, the arm's length GP/Sales of the comparable companies is 28.22 per cent which is lower than the GP/Sales of DCIL for the year ended 31-3-2003 of 61.10 per cent indicating that DCIL has retained more than the arm's length margin at the gross level. He stated that transfer pricing Legislation as provided in the Income-tax Act allows a taxpayer to have an option to compute the ALP which may vary from arithmetic mean by an amount not exceeding 5 per cent of such arithmetic mean. For this he quoted proviso to section 92C(2) of the Income-tax Act which states that "where more than one price is determined by the most appropriate method, the arm's length price shall be taken to be the arithmetical mean of such prices or at the option of the assessee, a price which may vary from the arithmetical mean by an amount not exceeding five per cent of such arithmetical mean." The assessee computed the arm's length price considering the 5 per cent tolerance range. The results of such computation are given below :—

Shri Rahul Mitra submitted that based on the above analysis, the amount of adjustment in the international transaction of the assessee with DCIL would be USD 275632. In case of assessment year 2004-05, the arm's length GP/Sales of the comparable companies is 28.75 per cent which is higher than the GP/Sales of DCIL for the year ended 31-3-2004 of 25.26 per cent indicating that DCIL has not retained more than the arm's length margin.

23. The learned DR could not refute the arguments and facts discussed by Shri Rahul Mitra. He only reiterated that the Commissioner (Appeals) during the course of appellate proceedings had argued against the analysis that DCIL was the main company and that the assessee was only a paper company and not a company of substance. Based on his findings from the internet the Commissioner (Appeals) had understood the assessee to be the main company of the Group and DCIL, he argued was a paper company created only to evade taxes. Hence, in his order he stated that there is no justification for allowing 28 per cent margins to DCIL and entire profits should come to DCPL."

Section 92CA(4) provides that upon receipt of the order of the TPO, the Assessing Officer shall compute the total income of the assessee, having regard to the arm's length price determined by the TPO. Relevant extracts of sub-section (4) of section 92CA are reproduced as under :

"(4) On receipt of the order under sub-section (3), the Assessing Officer shall proceed to compute the total income of the assessee under sub-section (3), the Assessing Officer shall proceed to compute the total income of the assessee under sub-section (4) of section 92C having regard to the arm's length price determined under sub-section (3) by the Transfer Pricing Officer."

As discussed above, upon receipt of the order of the TPO, the Assessing Officer would compute the total income of the assessee under sub-section (4) of section 92C Relevant extract of the said sub-section are reproduced as under :

"(4) Where an arm's length price is determined by the Assessing Officer under sub-section (3), the Assessing Officer may compute the total income of the assessee having regard to the arm's length price so determined...."

Based on the above, the learned counsel submitted that section 92CA(4) requires that the total income of the assessee is to be computed "having regard to" the arm's length price and not only either, based on, or in conformity with, the order passed by the TPO. Where the words 'having regard to' are used, the authorities should not mechanically consider only the factor which the statute had drawn their attention to, but consider all other reasonable aspects which could have a bearing on the matter. In this regard, as the expression 'having regard to' has not been defined in the Act, the counsel has drawn reference to the definition as provided in Law Lexicon. The definition of the expression 'having regard to' is as under :

"It indicates that in exercising the power, regard must be had also to the factors enumerated together with all factors relevant for exercise of that power."

Further, the Black's Law Dictionary defines the term 'regard' as 'consideration', 'attention' or 'care'.

5.52 Based on the above, the learned counsel submitted that computing the total income 'having regard to' the arm's length price does not simply mean that the price should be used as it were being used in a formula. On the contrary, it would mean that the arm's length price is one of the factors to be expressly considered while computing the total income. While so considering the factor, the Assessing Officer has to apply his judgment to the extent to which he should be influenced by the arm's length price. In this regard, the learned counsel placed reliance on the decision of the Privy Council in the case of CIT v. Williamson Diamonds Ltd. [1959] 35 ITR 290 . In the said decision, the Privy Council while analyzing the relevant of the expression 'having regard to' observed that 'all matters relevant to the question of unreasonableness should be considered.'

5.53 The learned counsel also submitted that the Assessing Officer simply went ahead and issued a notice to the assessee. This is further substantiated by the fact that the Assessing Officer issued a notice to the assessee within one day of receiving the TP order. It appears as though the Assessing Officer, being an ACIT, felt himself bound by the direction of the TPO, being an additional Director of Income-tax (i.e., an officer of a higher rank).

5.54 The learned counsel further submitted that the Assessing Officer neither took into consideration the submission filed by the assessee nor did the Assessing Officer grant an opportunity to the assessee of being heard. It is clear that the Assessing Officer simply went ahead and confirmed the adjustment proposed by the TPO, without 'having regard' to any facts or submissions made by the assessee. Accordingly, the learned counsel submitted that the process followed by the Assessing Officer for concluding the transfer pricing assessment was flawed.

5.55 In this regard, the CIT(A), in her order, has stated that the assessee made no additional submissions and simply reiterated the submissions made before the TPO. In this connection, it should be noted that on receipt of the show-cause notice from the Assessing Officer on 29-3-2006, the assessee filed submissions on 31-3-2006 at 5 PM (being the date on which the Assessing Officer passed the assessment order). Accordingly, it is a remote possibility that the Assessing Officer would have perused the submissions filed by the assessee and the income to a conclusion that the submissions were a repetition of the submissions filed with the TPO.

5.56 In relation to the second part of the question, the learned counsel submitted that the section 92C(3) provides that the Assessing Officer may proceed to determine the arm's length price if he is of the opinion that certain conditions have been satisfied. Relevant extracts of the said sub-section are reproduced as under :

"(3) Where during the course of any proceeding for the assessment of income, the Assessing Officer is, on the basis of material or information or document in his possession, of the opinion that—

(a) the price charged or paid in an international transaction has not been determined in accordance with sub-sections (1) (and (2); or
(b) any information and document relating to an international transactions have not been kept and maintained by the assessee in accordance with the provisions contained in sub-section (1) of section 92D and the rules made in this behalf; or
(c) the information or data used in computation of the arm's length price is not reliable or correct; or
(d) the assessee has failed to furnish, within the specified time, any information or document which he was required to furnish by a notice issued under sub-section (3) of section 92D.

the Assessing Officer may proceed to determine the arm's length price in relation to the said international transaction in accordance with sub-sections (1) and (2), on the basis of such material or information or document available with him...."

The above-mentioned section clearly provides that before the arm's length price is determined, the Assessing Officer has to prove that at least one of the four conditions laid down in sub-section (3) above have been satisfied. However, in the instant case, the Assessing Officer did not prove to the assessee that the above conditions were satisfied, either before initiating the transfer pricing assessment or during the course of the proceedings. The learned counsel submitted that, in its case, the Assessing Officer has failed to establish that any of these four conditions have been satisfied. In particular, there is nothing in the order of the Assessing Officer or of the TPO to suggest that clause (c) of section 92C(3) is satisfied.

5.57 The learned counsel drew reference to Circular No. 12/2001 dated 23-8-2001 issued by the CBDT, wherein the following has been mentioned:

"(iii)It should be made clear to the concerned Assessing Officers that where an international transaction has been put to a scrutiny, the Assessing Officer can have recourse to sub-section (3) of section 92C only under the circumstances enumerated in clauses (a) to (d) of that sub-section and in the event of material information or document in his possession on the basis of which an opinion can be formed that any such circumstance exists. In all other cases, the value of the international transaction should be accepted without further scrutiny."

Based on the above-mentioned para, it is clear that the intention of section 92C(3) has always been that scrutiny of the international transactions of an assessee can only be done if the Assessing Officer can prove that the circumstances enumerated in clauses (a) to (d) are satisfied. In the instant case, the TPO/Assessing Officer have at no stage of the assessment proceedings, provided any information or documentation which could establish that the procedure followed by the assessee or the computation of the arm's length price by the assessee was erroneous. The TPO/Assessing Officer simply conducted their analysis and applied the same to the assessee's case. This approach of the TPO/Assessing Officer clearly demonstrates bias. In this connection, reference is also drawn to Circular No. 14/2001 issued by the CBDT, which also provide that only if such deficiency or insufficiency is found in the documentation of the assessee, the TPO/Assessing Officer can conduct a scrutiny. The relevant extract of the circular has been reproduced below :

"Where such onus is discharged by the assessee and the data used for determining the arm's length price is reliable and correct, there can be no intervention by the Assessing Officer. This is made clear by sub-section (3) of section 92C which provides that the Assessing Officer may intervene only if he is, on the basis of material or information or document in his possession, of the opinion that the price charged in the international transaction has not been determined in accordance with sub-sections (1) and (2), or information and documents relating to the international transactions have not been kept and maintained by the assessee in accordance with the provisions contained in sub-section (1) of section 92D and the rule made thereunder; or the information or data used in computation of the arm's length price is not reliable or correct...."

Further, even in the case of Mentor Graphics (Nokia) (P.) Ltd. (supra), the Tribunal, at Para 39.4 of the order, have observed as follows :-

"The TPO could have carried fresh search only of the comparable drawn by the taxpayer was insufficient or had other deficiency."

5.58 The learned counsel had drawn attention to the Tribunal's additional analysis which also substantiates that the transfer pricing of the assessee with its associated enterprises is at arm's length.

5.59 Rule 10B(1)(e) lays down the manner in which TNMM is to be applied for computing the arm's length price. Sub-clause (i) of said Rule provides that the net profit margin realized by the assessee can be computed in relation to the following :

(a) costs incurred,
(b) sales effected,
(c) assets employed (or to be employed); or
(d) any other relevant base.

The TP Study as well as the TPO has computed the arm's length price under TNMM using 'costs incurred' as the relevant base. If one were to compare the net margins using assets employed (i.e., capital employed) as the relevant base, the following scenario would emerge :—

PBIT: Net profit before interest and tax.

Capital employed : Shareholders fund and debt fund as at the beginning of the reporting period. The above table clearly demonstrates that the international transactions of the assessee satisfy the arm's length test on the basis of comparison of return on assets.

If one were to run a fresh search for comparable companies for the relevant period, the results would be as under :

The above search was conducted using PROWESS database (i.e., the same database used by the TPO), as updated on 15-2-2008, using the following economic activities :

(a) Computer software; and
(b) Software services and consultancy.

No additional quantitative/quantitative filters were applied.

The derived man-hour billing rate of the assessee works out USD 29.24. This rate has been computed after taking into account the turnover of the assessee, billable man-hours during the financial year 2002-03 and IN/ USD rate of 48.27 (which is the average of the rate on 1-4-2002 and 31-3-2003). Further, the TPO in the assessment made for assessment year 2002-03 has herself used an industry benchmark of USD 18-25 per hour. If the industry rates are considered as a potential CUP, the man-hour rate of the assessee and the consequent value of the international transactions of the assessee with its associated enterprises would be at arm's length. Based on the above, three additional analysis/workings, the learned counsel submitted that the international transactions of the assessee with its associated enterprises during assessment year 2003-04 are at arm's length.

5.60 In case of a reference being made to the TPO, the TPO passes his order under section 92CA(3), a copy of which is forwarded to the Assessing Officer. Section 92CA(4) (as it stood prior to the amendment made by the Finance Act, 2007) provides that upon receipt of the order of the TPO, the Assessing Officer shall compute the total income of the assessee 'having regard' to the arm's length price determined by the TPO.

"(4) On receipt of the order under sub-section (3), the Assessing Officer shall proceed to compute the total income of the assessee under sub-section (4) of section 92C having regard to the arm's length price determined under sub-section (3) by the Transfer Pricing Officer."

The Finance Act, 2007, has amended section 92CA (4) to replace the words 'having regard to' with 'in conformity with'. This amendment has been made effective from 1-6-2007. Since the amendment is effective only from 1-6-2007, it would not be relevant for assessment year 2003-04, and accordingly, the provisions prior to the amendment would apply in the instant case.

5.61 The CIT(A) in her order, has stated that the provisions relating to 'having regard to' in section 92CA(4) was a lacuna in the statute which has now been rectified by the Finance Act, 2007. In this regard, as mentioned above, the amendment made to this section has been made effective only from 1-6-2007 and not before. The legislature in its wisdom has made the amendment effective from 1-6-2007, and it would not be appropriate for the CIT(A) to consider the amended provisions as being applicable for periods prior to 1-6-2007. The learned counsel placed reliance on the decision of the Tribunal in case of Aztec Software & Technology Services Ltd. (supra), which held that the amendment made by the Finance Act, 2007 is applicable from 1-6-2007. Relevant extracts of the decision are reproduced as under :

"53. Now words 'having regard to' have been replaced by words 'in conformity with'. So now Assessing Officer after introduction of sub-section (4) above is required to pass assessment order in conformity with the order of the TPO determining arm's length price. Now the order of the TPO has been expressly made binding on the Assessing Officer."

5.62 In the transfer pricing assessment made on the assessee for assess-ment year 2002-03 after doing a detailed scrutiny, the TPO had upheld the international transactions of the assessee with its associated enterprises to be at arm's length. For this purpose, the TPO had compared the derived off-shore billing rates of the assessee with the comparable off-shore rates (as Comparable Uncontrolled Price or CUP) as available on transactions of the assessee were at arm's length, was also acknowledged by the Assessing Officer. Accordingly, the Assessing Officer did not make any transfer pricing adjustment in assessment year 2002-03. However, during the transfer pricing assessment for assessment year 2003-04, the TPO has taken a completely different view and has concluded the assessment considering the TNMM as the most appropriate method, disregarding the use of the CUP method as used during the TP assessment for assessment year 2002-03.

5.63 In this regard, the ld. Counsel submitted that the principles of res judicata, under which, amongst other reasons, unless new facts are brought on record, the conclusion arrived at during the assessment in respect of one year, cannot differ from the assessment in respect of the subsequent year. Further, various High Courts, have held that when a question of law or fact is decided in the assessee's own case for an earlier, the Tribunal will be justified in placing reliance on the earlier decision to base its conclusion, in the absence of any new material or change in circumstances or a fresh look necessitated on existing facts on a closer and more intelligent analysis. Reliance in this regard is placed by the learned counsel on the following decisions :

(a) Commissioner Of Income Tax v. M/S. Velimalai Rubber Co. Ltd. [1991] 181 ITR 299 (Ker.)
(b) Annamalai Reddiar…Applicant;* v. Commissioner Of Income-Tax, Kerala…. [1964] 53 ITR 601 (Ker.).

The learned counsel further submitted that in the context of transfer pricing, there is normally a presumption that the principle of res judicata does apply. This is clear from the proviso to Rule 10D(4), which reads as below :

"Provided that where an international transaction continues to have effect over more than the previous year, unless there is any significant change in the nature or terms of the international transaction, in the assumption made, or any other factor which could influence the transfer price, and in the case of significant change, fresh documentation as may be necessary under sub-rules (1) and (2) shall be maintained bringing out the impact of the change on the pricing of the international transaction."

The above rule clearly stipulates that so long as there is no significant change in the nature/terms of international transactions, there is no requirement for creating fresh set of documentation for the subsequent year.

As a corollary,

If :

(a) The value of the international transactions as recorded in the books of account has been accepted during an earlier year as the arm's length price; and
(b) There is no significant change in the nature/terms of international transactions in the subsequent year.

Then :

there is no reason to reject the value of the international transactions as recorded in the books of account for the subsequent year.

5.64 The learned counsel had further argued that even in the absence of the specific provisions of Rule 10D(4), the general rule that the principle of res judicata is not applicable to the decision of tax authorities, has certain clear limitations.

5.65 If a given case falls within the parameters of the following limitations explained in H.A. Shah & Co. v. CIT& Excess Profits Tax [1956] 30 ITR 618 (Bom.) and numerous other decisions, the tax authority would not be entitled to unsettle the earlier findings :

(i) Where the earlier finding was not arbitrary or perverse;
(ii) Where the earlier finding was arrived at after making due inquiries;
(iii) Where no fresh facts were placed before the authority giving the later decision;
(iv) Where the authority giving the earlier decision had taken into consideration all material evidence; or
(v) Where unsettling the earlier decision leads to injustice to the assessee.

5.66 The learned counsel had submitted that its case falls within each of the limitations enumerated above. Based on a detailed inquiry/analysis, the Assessing Officer/TPO reached a conclusion in assessment year 2002-03 that the pricing of the assessee was at arm's length. Further, all the facts pertaining to assessment year 2002-03 were similar to the facts relating to assessment year 2003-04, such as :

- Nature of services i.e., provision of software development services;
- Contractual terms; and
- Method of receiving remuneration.

Accordingly, the learned counsel submitted that the principles of res judicata should apply in its case, and it is completely inappropriate on the part of the Assessing Officer/TPO to have reached a different conclusion in assessment year 2003-04.

5.67 The arguments of the ld. DR are therefore factually incorrect. The order for assessment year 2003-04 was passed by the TPO and Assessing Officer in March, 2006. It was only based on the conclusions reached in the proceedings of this year that the CIT subsequently revised the order passed for assessment year 2002-03 The revision order for assessment year 2002-03 was passed in March, 2007.

5.68 The decision of the ITAT, Delhi in Mentor Graphics (Nokia) (P.) Ltd.'s case (supra) is the only decision which specifically deals with a captive software development service provider. However, the CIT(A) has disregarded the decision of the Tribunal. In this regard, the learned counsel provided below, the commonality of the facts between the assessee and Mentor Graphics, being the assessee in the cited case.

The Tribunal, in the said case, has inter alia laid down the following guiding principles :

(a) A TPO can reject comparables in the TP Study of the taxpayer only if the comparables were insufficient or deficient, (refer to the discussions in issue No. 10.1).
(b) Methodology of selection appropriate comparable companies is dependent on the principle of FAR analysis ( i.e., functions performed, assets employed and risks assumed) (refer to the discussions in issue Nos. 10.2 and 12).
(c) Comparable companies having related party transactions is against the basics of transfer pricing, and therefore ought not to be considered for the purpose of bench marking (refer to the discussions in issue No. 10.3)
(d) Adjustment for normal open market risks which the taxpayer was insulated from, being a captive service provider should be made. Some adjustments are adjustments for working capital, risk and growth and research and development expenses, etc. (refer to the discussions in issue No. 10.4 and 11).

5.69 The learned counsel submitted that the CIT(A) ought to have followed the guidelines issued by the Delhi Tribunal and provided appropriate relief to the assessee, as the said decision is squarely applicable in the case of the assessee. The following table shown some of the similarities between the assessee and E-gain Communication, being the assessee in the case of E-gain Communication (P.) Ltd. (supra) :

5.70 The learned counsel submitted that all the six peculiarities mentioned by the DR are equally applicable in the instant case :

(a) Peculiarity No. 1 : The TPO has also used data for other periods for the purpose of his analysis.
(b) Peculiarity No. 2 : The TPO has also used contradictory approaches for conducting his analysis.
(c) Peculiarity No. 3 : The TPO has not looked into the FAR analysis of the companies considered by him as comparables.
(d) Peculiarity No. 4 : The rejection of the comparables in the TP Study is without proper justification.
(e) Peculiarity No. 5 : The TPO has considered companies having a wide range of profit margins (particularly high profit margin cases).
(f) Peculiarity No. 6 : Inconsistencies are noted in TPOs selection/rejection even after applying his own filters.

Accordingly, the decision in the case of Mentor Graphics (Nokia) (P.) Ltd. (supra) is squarely applicable in the instant case. As discussed, the contention of the DR that the assessees bears pricing risk is without any merits. Further, during the course of the hearing on 25-8-2008, the DR filed a written note highlighting some of the issues arising out of the decision of the Tribunal in the case of Mentor Graphics (Nokia) (P.) Ltd. (supra). The learned counsel pointed out that the DR has quoted the observations of the Tribunal in erroneous contexts, and the note is another desperate attempt by the DR to confuse the matter at hand. In any case, the assessee has appropriately addressed the relevant arguments made by the DR at the appropriate places in this note.

5.71 We, therefore, summarise our conclusion as follows :

(i) Since the basic intention behind introducing the transfer pricing provisions in the Act is to prevent shifting of profits outside India, and the assessee is claiming benefit under section 10A of the Act, the transfer pricing provisions ought not to be applied to the assessee.
(ii) Circular No. 14/2001 issued by the CBDT is binding upon the TPO.
(iii) There was no infirmity in the TP Study conducted by the assessee, and the TPO erred in disregarding the same for the purpose of computing framing the assessment and making the transfer pricing adjustment.
(iv) The TPO or the Assessing Officer needs to satisfy and communicate to the taxpayer the relevant clause under section 92C(3) which has been triggered by the assessee, which has necessitated the application of provisions of the transfer pricing provisions, in the instant case, since this was not demonstrated to the assessee, the transfer pricing order is void.
(v) The TPO erred in conducting a fresh study for the purpose of passing his order. The study conducted by the TPO is not in conformity with the provisions of Rules 10B(4) and 10D(4).
(vi) The TPO erred in disregarding the most appropriate method adopted by the assessee in the TP Study, and also in using the Prowess data base. The TPO did not provide any reason for deviating from the TP Study in respect of these matters.
(vii) The TP Study cannot be ignored by the TPO, in the absence of any deficiency or insufficiency. Further, the order passed by the TPO appears to have been passed with the intention of making a higher transfer pricing adjustment.
(viii) For the purpose of comparability, companies with even a single rupee of transactions with associated enterprises cannot be considered as compatibles.
(ix) Adjustment needs to be made to the margins of the comparables to eliminate differences on account of different functions, assets and risks. More specifically, adjustment needs to be made for :

(a) Differences in risk profile.
(b) Difference in working capital position.
(c) Differences in accounting policies.

(x) The TPO has grossly erred in 'normalising' the profits of super profit companies. Such companies should have been excluded from the list of comparables.
(xi) The proviso to section 92C(2) of the Act provides a standard deduction of 5 per cent to the taxpayers. The only condition for availing this benefit is that it is subject to the option of the taxpayer.
(xii) The decision of the Tribunal in the case of Mentor Graphics (Nokia) (P.) Ltd. ( supra) and E-gain Communication (P.) Ltd.'s case ( supra) are squarely applicable to the assessee's case. Some of the relevant common facts relating to the assessee before us and the assessee's in both the decisions mentioned above have been summarised below :

(xiii) Based on the issues raised and discussed, it should be concluded that the transactions of the assessee with its associated enterprises satisfy the arm's length test, and that the order of the TPO is bad in law and on facts.
(xiv) Without prejudice to the submission of the assessee that the comparables selected by the TPO should not be considered for the purpose of comparability analysis, the assessee has prepared a working carrying out an accept/reject test on the comparables of the TP Study as well as the companies selected by the TPO as comparables. Even on the basis of this statement, the transactions of the assessee with its associated enterprises satisfy the arm's length test.

In the result, the appeal is allowed.