Pramod Kumar AM:-1. These two appeals, which pertain to the same assesse and involve some common as well as interconnected issues, will be taken be taken up together. In ITA No. 3223/Ahd/11, the assesse appellant has called into question correctness of the order dated 24th October 2011 passed by the Assessing Officer under section 143(3) r.w.s. 144C (13) of the Income Tax Act, 1961, for the assessment year 2007-08. In ITA 2858/Ahd/12, the order challenged in this appeal is dated 22th October 2012 passed by the Assessing Officer under the same statutory provisions, for the assessment year 2007-08.

2. The short issue that we really need to adjudicate in these cases is as to what is the most appropriate method, on the facts of this case, for determining the arm’s length price of the exports made by the assessee to its associated enterprises (AEs) abroad.

3. While the assessee has contended that the transactional net margin method (TNMM) is the most appropriate method on the facts of this case, the stand of the authorities below is Comparable Uncontrolled Price (CUP) method is the most appropriate method for ascertaining the arm’s length price for the assessment year 2007-08, while Cost Plus Method (CPM) is the most appropriate method for ascertaining the arm’s length price for the assessment year 2008-09.

4. To adjudicate on these appeals, only a few material facts need to be taken note of. The assessee before us is engaged in the business of manufacturing glass mosaic products and selling the same to its AEs. During the course of the proceedings before the TPO for the assessment year 2007-08, it was noted that the assessee has used transactional net margin method to determine the arm’s length price. It was also noted that the companies selected by the assessee as comparables, for the purposes of TNMM, were manufacturing glass bottles, kitchen glass wares, insulated glass, laminated glass etc but then these products are completely different from the glass mosaic products manufactured by the assessee. It was also noted that ‘glass mosaic is a luxury product having technical and aesthetic characteristics much different from a glass product manufactured by other companies”. It was also noted that the main raw material in the glass industries is Silica but it constitutes less than 3% of the cost of production. The TPO, therefore, proceed to reject the TNMM on the ground that “comparable data is not available”. An effort was then made by seeking data on the cost analysis of the product sold to AEs and non-AEs so as to apply the internal cost plus method but this effort of the TPO had to aborted for want of relevant data. It was in this background that the TPO proceeded to adopt the CUP method. It was noted that the assessee has, in the relevant previous year, sold the same products to Palladio Glass Pvt Ltd (Rs 14,27,786) and Pino Bisazza Glass Pvt Ltd (Rs 2,15,13,437). Both of these entities are admittedly group companies of the assessee inasmuch as ultimate parent company of the assessee, i.e. Trend Group SpA, holds 33% equity in Palladio Glass Pvt Ltd, 67% equity in Pino Bisazza Glass Pvt Ltd an 99% equity in Gemstone Glass Pvt Ltd, i.e. this assessee. The assessee’s contention that the prices at which it has sold goods to these group entities cannot be treated as valid comparable for CUP method inasmuch such prices must be in respect of the transactions with unrelated parties, whereas these entities were admitted related parties, was rejected on the ground that “the transaction of the assessee was with resident Indian companies and there is no motive for tax avoidance”. While adopting the CUP method, the TPO referred to assessee’s stand to the effect that “there are no major players in similar decorative glass products in India.....(and).. there are limitations on availability of data pertaining to the similar decorative product industry” , and justified adoption of CUP method on the ground that transactional net margin method cannot be applied to the facts of this case for want of suitable comparables. He also justified his action on the ground that cost plus method cannot be applied on the facts of this case as the assessee did not give costing of the products sold to the AEs and the non AEs. As for the assessee’s contention that prices at which products are sold in India could not be compared with the prices at which the same products are sold abroad, on account of geographical differences, the TPO rejected the same by observing that “it is seen that the assessee company has sold its products at lower rate to the European market” but the “European market (in which AEs are based) is a premium market for luxurious good like the decorative products manufactured by the assessee and rate of these products should be higher than the (rates prevailing in) Indian market”. On the basis of this reasoning and adopting the price at which the same products are sold to group entities in India as the arm’s length price, the TPO recommended a transfer pricing adjustment of Rs. 5,09,77,309. Aggrieved by the arm’s length price adjustment so proposed, assessee approached the Dispute Resolution Panel. The DRP rejected the grievances of the assessee by observing, inter alia, that “the TP method adopted by the assessee was not more appropriate than adopted by the TPO” and that but for the adjustment for geographical differences, “CUP is more appropriate method and cannot be faulted with”. Accordingly, the ALP adjustment of Rs. 5,09,77,309 was upheld. The assessee is aggrieved and is in further appeal before us. As for the assessment year 2008-09, the TPO noted the contention of the assessee that the TNMM is most appropriate method for determining the ALP. He, however, also noted that in the transfer pricing report filed by the assessee it has been accepted that “the cost plus method is the most appropriate method when the business activity is in the nature of contract manufacturing”. It was also noted that in the FAR (function, asset and risk) analysis of the assessee has been characterized as contract manufacturer. The Transfer Pricing Officer further noted that “the assessee did not choose this method as the most appropriate method since the data corresponding to the normal gross mark ups, as required for the application of cost plus method, was not available in the public domain.” He then added that “it may be pertinent to mention here that in its reply to the show cause notice, the assessee did not raise any contentions about inappropriateness of CPM as the most appropriate method”. It was in this backdrop and relying upon the provisions of Section 92C(3) of the Act and the judicial precedents in the Coca Cola Inc Vs ACIT [(2009) 177 Taxmann 103 (P&H)], SAP Labs India Pvt Ltd Vs ACIT [(2011) 44 SOT 156 (Bang)] and Serdia Pharmaceuticals India Pvt Ltd Vs ACIT [(2011) 44SOT 391(Mum)], the TPO proceeded to reject the TNMM and adopt CPM for determining the arm’s length price. While doing so and accepting the position that the requisite data was not available at the point of time when the transfer pricing documentation was prepared, the TPO also observed that “the non availability of data at the time of preparation of the transfer pricing document need not fetter the application of correct method”. It was in this backdrop and after obtaining the requisite information under section 133(6) that the TPO proceeded to make an ALP adjustment of Rs. 2,14,46,604. Once again, aggrieved, inter alia, by this change of method of ascertaining the arm’s length price, the assessee carried the grievances before the DRP. Once again, DRP rejected the grievances of the assessee. While doing so, the DRP, inter alia, observed that “.....We must concede that perfect data for comparability analysis is rarely available (irrespective of) whichever method is selected. In such circumstances, the assessee’s argument that CPM should be rejected because perfect data is not available does not have much force”. The assessee is not satisfied with the stand so taken by the DRP, and is in appeal before us.

5. We have heard the rival contentions, carefully perused the material on record and duly considered facts of the case in the light of the applicable legal position.

6. We have noted that so far as assessment year 2007-08 is concerned, the TPO has applied internal CUP method to ascertain the arm’s length price. It is only elementary that comparable uncontrolled price at which the entity has sold the same product to an independent enterprises is a sine qua non for application of internal CUP. It is so for the reason that Rule 10 B(1)(a) provides that under comparable uncontrolled price method, as a first step, the price charged or paid for property transferred or services provided in a comparable uncontrolled transaction, or a number of such transactions, is identified. Rule 10A(a), on the other hand, defines ‘uncontrolled transaction’ “a transaction between enterprises other than associated enterprises, whether resident or non-resident”. Therefore, the first essential input for application of CUP method is the price charged or paid for similar product in a transaction between two enterprises, whether resident or non-resident, which are not associated enterprises. In the present case, the comparable price adopted for determining the arm’s length price is the price at which the assessee has sold the same product to other group entities, which are thus ‘associated enterprises’, residents in India. It has been defended by the DRP on the ground that there cannot be any tax avoidance motive in selling the products at an artificial price. In our considered view, however, this aspect of the matter is irrelevant inasmuch as the very definition of ‘uncontrolled transaction’ under rule 10A excludes the transactions with associated enterprises “whether resident or non-resident”. Once it is not in dispute that “uncontrolled transaction” is a statutorily defined term, there is no room for discarding or questioning this definition on the basis of superior logic in an alternative definition. Such heroics are not called for in the process of judicial interpretation. Learned DRP ought to have followed the law as it exists rather than pondering over what the law ought to be, as is inherent in their justification for inclusion of uncontrolled transactions with associated enterprises resident in India. We disapprove the line of reasoning adopted by the authorities below. In our considered view, whether the transactions are with associated enterprises resident in India or with associated enterprises resident outside India, the prices at which such transactions are entered into with such enterprises cannot be taken as “comparable uncontrolled price” for the purpose of determining the arm’s length price.

7. As for the question as to whether the transactions with associated enterprises can, in any situations, be considered as a valid input for ascertainment of arm’s length price under the CUP method, we find the issue is covered, in favour of the assessee, by decision of a coordinate bench in the case of Sabic Innovative Plastic India Pvt Ltd Vs DCIT [(2013) 59 SOT 138 (Ahd)]. In this case, it was inter alia observed that:

7. As for the connotations of the expressions ‘Internal CUP' and ‘External CUP', while the former deals with a situation in which prices at which similar transaction is entered into with an independent enterprises, latter refers to the situation in which two unconnected independent enterprise deal with each other. In other words, even in the case of ‘internal CUP', the arm's length price to be adopted is the price, subject to admissible adjustments, at which the similar transactions are carried out between the assessee and an independent enterprise. Internal CUP has nothing to do with the margins earned by the same enterprises from other transactions, as is the case before us. Learned CIT(A)'s reliance on the decision of a coordinate bench of this Tribunal, in the case of Bayer Material Science Pvt Ltd Vs Additional Commissioner of Income Tax (134 ITD 582), is also of no longer sustainable in law. The authorities below have proceeded on the basis that in the case of GEII's plastic division there were substantial intra AE transactions as evident from learned CIT(A)'s observations to the effect that, “the appellant has also stated that for the purposes of internal CUP also, the uncontrolled transactions can only be considered. Since, the transactions of the predecessor company i.e. GE India Pvt. Ltd. were not uncontrolled transactions, but were transactions made with Associated Enterprises under controlled conditions……". Yet these controlled transactions were treated as good comparables. Bayer Material Science was also a case in which transactions between associated enterprises were used as comparables for benchmarking, and this action was upheld by the Tribunal. It was in this process that in the case of Bayer Material Science (supra), a coordinate bench of this Tribunal had inter alia observed as follows:

19. It is possible that the nature of international transaction between two associated enterprises may be such which, in normal course, is unusual between independent enterprises. In such a case there will be hardly any comparable uncontrolled case for the purposes of benchmarking of such transaction. The question will arise as to whether in such a situation, the transfer pricing provisions will fail and cease to be applicable and as such the TPO will be compelled to accept the manoeuvred price declared by the assessee. The further question will be as to whether any cognizance can be taken of such controlled transactions for benchmarking. We have observed above that a majority of assesses do not intend to play foul with the Revenue by unnecessarily attempting to reduce the tax liability. In such circumstances the declared income from such international transactions will itself represent the arm's length price. Thus, where it is an admitted position between the tax payer and the tax collector that there is no comparable uncontrolled transaction due to the nature of transaction being such that it is ordinarily between associated enterprises, in such a case, a transaction between two associated enterprises at arm's length price, though technically called 'controlled transaction', would partake of the character of 'uncontrolled transaction' for the purposes of determining the ALP in a later international transaction between two AEs. In such a situation, no fetters can be placed on the powers of the TPO to consider such comparable controlled transaction - having adorned the garb of uncontrolled transaction -for the purposes of benchmarking. If the contention of the ld. A.R. is accepted that controlled transaction should be altogether shunted out for the purpose of transfer pricing provision, even in rare circumstances as are presently prevailing, then the very rationale and purpose of sections 92 to 92F, being the determination of ALP, which is otherwise achieved from the controlled transaction, will be defeated. It is in such exceptional circumstances that the principle of purposive interpretation will come into play to set free the hands of the TPO tied with determining ALP only on the basis of uncontrolled transactions.

20. We have noticed above that the purpose behind these provisions is to prevent the avoidance of tax in the international transactions by ascertaining the arm's length price. These provisions are basically for the assistance of the Revenue as is evident from sec. 92(3) which mandates that the provisions ofthis section shall not apply in a case where the computation of income under subsection (1) or the determination of allowance for any expense or interest under that sub-section or determination of any cost or expense allocated or apportioned has the effect of reducing the income chargeable to tax computed on the basis of entries made in the books of account in respect of the previous year in which the international transaction was entered into. The effect of section 92(3) is that if the determination of income from international transaction at arm's length price results into a lower income than what has been declared by the assessee as per the entries in the books of account, then no cognizance should be taken of such determination of ALP, which shall be ignored and the income shall be computed on the basis of entries made. On the other hand, if the income determined at arm's length price is higher than that emanating from entries in the books of account, then such income at arm's length price, being higher than that from the entries in the books of account, shall be included in the total income of the assessee. It is, therefore, manifest that the higher of income determined at arm's length price or as emerging from the entries made in the books of account, is taken into consideration for computing the total income of an assessee. This sub-section (3) of section 92 when seen in juxtaposition to the Chapter X in which the relevant sections have been resided titled as 'Provisions relating to avoidance of tax', makes it apparent that the purpose behind such provisions is to uncover the arrangement made by the associated enterprises in not reflecting the true profit from the international transactions. If we accept the contention raised by the ld. A.R. that the controlled transactions should be completely ignored in such a situation when there are no uncontrolled transactions at all, it would amount to defeating the object of these provisions. When the very purpose of these provisions is to determine arm's length price and there is admittedly no record of any uncontrolled transaction, in our considered opinion, it is perfectly in order to consider a controlled transaction genuinely entered in an uncontrolled manner between some other associated enterprises, for the purposes of benchmarking of such a transaction.

8. However, the same learned Member, as a Third Member in the case of Technimont ICB Pvt Ltd Vs Additional Commissioner of Income Tax (138 ITD 23 TM), and agreeing with one of us (i.e. the Accountant Member) in a situation of divergence of opinion between the members constituting the division bench, discarded the above school of thought. Learned Departmental Representative, however, urges us to follow the view taken by the Bayer Material Science (supra) and ignore the Third Member decision in the case of Technimont ICB (supra). In a written note filed before us, it is submitted that "In this regard, it is important to note that in the Third Member decision in Technimont ICB India (P) Ltd. case, curiously no discussion is made onthe earlier judgments delivered in the cases of NGC Network and Bayer Material Science. This assumes significance in this case because the Hon'ble member who wrote the Third Member judgment in Technimont case had earlier written the judgment in the case of Bayer Material Sciences also. The Hon'ble Member categorically wrote that no cognizance of the judgments cited by the representatives (including Bayer Material Sciences and NGC) are being made by him in the case (Technimont), which is an inherent contradiction in the stands taken by the Hon'ble Member in these cases. In such a scenario, the case should have been referred to the Special Bench instead of deciding the same issue in two different ways that too without distinguishing the earlier judgment…….Consequently, the extreme view taken in the Technimont judgment is incorrect, impractical and against the spirit of several judgments delivered by different Benches of Tribunals on this issue". We are not swayed by these submissions, and see no legally sustainable merits in the same. As far as the question of binding nature of these judgments is concerned, there does not seem to be much dispute on the proposition that a Third Member decision overrides the decision of a division bench and has a greater binding force. It has the same precedence value as that of a special bench. Elaborating this principle, a special bench of this Tribunal, in the case of DCIT Vs Oman International Bank SAOG (100 ITD SB 285), that a Third Member decision is defacto a decision of larger bench, and , in coming to this conclusion, the Special Bench was guided by Hon'ble Delhi High Court's judgment in the case of P C Purivs CIT (584) wherein Their Lordships observed that "There is no difference, really speaking, between a Full Bench of three Judges sitting together and this method of referring to the third Judge in the case of a difference of opinion between the two Judges. Whether the first method is adopted or the second, "opinion of the majority" will be decisive. In this case, there is a formal reference to a third Judge to ascertain his opinion. He is the deciding voice. He turns the scales. The third Judge is the Full Bench. Not alone. But along with the two others, who first heard the case. Whether the three Judges sit at the same time or at different times - two at one time and the third hearing the matter later on a difference of opinion - does not make much difference.". Viewed thus, a Third Member may not really be bound by the decisions of division benches and the Third Member decision may be seen as overruling not only the dissenting views but also unanimous decisions of the division benches. Learned Third Member, in his majority view in the case of Technimont (supra), had stated as follows:

14. ..................There is no statutory sanction for roping in a comparable controlled transaction for the purposes of benchmarking. When it has been clearly mandated in all the relevant methods for determining ALP that the comparison has to be made by the enterprise's international transaction with comparable uncontrolled transaction, by no sheer logic a comparable controlled transaction can be employed for the purposes of making comparison. There is no warrant for diluting the prescription given by the statute or rules when such prescription itself serves the ends of justice properly and is infallible. If the view of the Revenue that a controlled transaction should not be shunted out for the purposes of benchmarking, is accepted, then all the relevant provisions contained in Chapter X in this regard, will become otiose. If such a contention of making comparison with a comparable controlled transaction is taken to its logical conclusion, then there will never arise any need to take up any case for transfer pricing scrutiny. The reason is obvious. ALP is determined for application in respect of transactions between two AEs so that the profit likely to arise from such transactions is not under-reported vis-a-vis from similar transactions with third parties. If the comparison is made again with net profit margin realized from transactions between two AEs, instead of third parties, it may demonstrate the same cooked results in both the situations, thereby leaving no scope for any adjustment. In this eventuality, the very object of such provisions will be frustrated. Thus it follows that the ALP can be determined only by making comparison with a comparable uncontrolled transaction and not a comparable controlled transaction.

9. Learned Third Member thus virtually overruled his own view in a division bench. As he did so, he was not oblivious of the division bench order that he was departing from, as evident from his following closing remarks:

18. Before parting with this matter, I consider it my duty to record that the ld. AR relied on certain decisions including UCB India (P.) Ltd. v..Asstt. CIT [2009] 121 ITD 131 , Bayer Material Science (P.) Ltd. v. Addl. CIT [2012] 134 ITD 582 and Dy. CIT v. BP India Services (P.) Ltd. [2011] 133 ITD 255 (Mum) in which it has been held that controlled transactions cannot be considered for determining ALP in other transactions. Per contra , the ld. DR has relied on a solitary decision rendered by the Mumbai bench of the tribunal in NGC Network (India) (P.) Ltd. (supra) to buttress his contention that a controlled transaction can also be considered for benchmarking. I do not propose to embark upon these cases separately for discussion, I clarify that my decision in the foregoing paras is founded on the interpretation of the relevant bare provisions of the Act and Rules, without taking any assistance from decisions cited by the rival parties on the point, which differ in their conclusion as stated by the ld. Representatives before me.

10. Learned Departmental Representative's submission that a judicial officer cannot deviate from his own stand does not seem to be correct. What may be material is the hierarchical position of the forum at which the judicial officer is placed and not the judicial officer himself. A Third Member decision, as we have noted above in the light of Hon'ble Delhi High Court's judgment in the case of P C Puri(supra), is a de facto full bench decision and a view does seem possible that it is not, therefore, bound by division bench decisions- including, of course, the orders the Third Member himself may have authored on behalf of the division benches. In any case, no one is infallible and there cannot be any heroism in perpetuating an error either. Banjamin N Cardozo, one of the most distinguished US Chief Justices, in his classic book ‘The Nature of Judicial Process' (Yale University Press; first published in Dec 1921; online version at http://www.constitution.org/cmt/cardozo/jud proc.htm) beautifully puts it as “ I own that it is a good deal of a mystery to me how judges, of all persons in the world, should put their faith in dicta. A brief experience on the bench was enough to reveal to me all sorts of cracks and crevices and loopholes in my own opinions when picked up a few months after delivery, and reread with due contrition. The per-suasion that one's own infallibility is a myth leads by easy stages and with somewhat greater satisfaction to a refusal to ascribe infallibility to others." The hyper technical issues raised by the learned Departmental Representative are thus devoid of any legally sustainable merits. The issue stands covered against the revenue by a Third Member decision which prevails over a division bench decision. Learned counsel submits that even if Bayer decision (supra) is to be implemented on merits, the assessee's case will not be covered. That aspect of the matter is, however, academic now. The course of action adopted by the authorities below, in adopting the net margin @ 8.95%, on the basis of results shown by the GEII's plastic division, is unsustainable in law, and we vacate the same. However, we have noted that the learned CIT(A) has declined to deal with all other issues on the short ground that the decision to apply, what he erroneously terms as, ‘Internal CUP' method for ascertaining arm's length price is correct and does not call for any interference. This stand, in the light of the discussions above, is clearly incorrect and unsustainable in law. In this view of the matter, as also bearing in mind entirety of the case, we deem it fit and proper to remit the matter to the file of the CIT(A) for fresh adjudication on merits in the light of above observations, in accordance with the law and by way of a speaking order. There cannot be any good reasons for us to deal with all other issues on merits when learned CIT(A) did not have an occasion to deal with the same at all. Learned CIT(A) has decided the matter on the short legal ground, which we find to be unsustainable in law, and, therefore, we send the matter back to the CIT(A) for proper adjudication. With these observations, matter stands restored to the file of the CIT(A) and all other grounds of appeal stand dismissed as infructuous.

8. It is, therefore, clear that the prices on which the assessee has sold the same products to resident associated enterprises cannot be taken as benchmark for ascertaining the arm’s length price of its similar sale transaction with the non-resident enterprises. Once we come to the conclusion that necessary inputs for ascertaining the ALP under CUP are not available, there cannot be any occasion to apply the same. In the assessment year 2008-09, the TPO himself abandoned the CUP method and resorted to cost plus method.

9. As for the cost plus method being applied to the facts of this case, we have noted that the DRP has conceded that appropriate data for the application of cost plus method is not available and yet they have approved adopted of the CPM because, according to the DRP, appropriate data is not available for the TNMM either. It is difficult to understand the logic of this approach. CPM is not a residuary method in the sense that if every other method of ascertaining the arm’s length price fails, CPM can be applied on the basis of imperfect data. If at all there is a residuary method, or what is termed as the method of last resort, it is transactional net margin method. TNMM has almost become the ‘default’ method for taxpayers in recent years.The key advantage of the TNMM is that there is often available data in the public domain about the net profits that comparable independent businesses earn from their trading activities in comparable markets with other third parties. As such, the TNMM often proves easier to apply than, say, the Cost Plus or RPM methods, and TNMM is less sensitive to minor differences in the products being sold. It is also important that the data for application of CPM, as collected by the TPO, was not in public domain and this has been done by collecting information under section 133(6). Such a data could not have been available to the assessee. All shortcomings being equal in the application of the methods of determining the ALP, as at best is the case of the authorities below, the method to be preferred is the method for which necessary inputs are available in the public domain. As for the TNMM data not being available, as there is difference between the product that the assessee is manufacturing vis-à-vis the products being manufactured by the comparables adopted, it is only broad similarity in the product and economic similarity in the conditions which is need. While on this issue, it may be relevant to refer to following observations in the UN’s Transfer Pricing Manual:

TNMM is usually applied with respect to broad comparable functions rather than particular controlled transactions. Returns to these functions are typically measured by a PLI in the form of a net margin that arguably will be affected by factors unrelated to arm’s length pricing. Consequently, one might expect the TNMM to be a relatively disfavoured method. Nevertheless TNMM is typically applied when two related parties engage in a continuing series of transactions and one of the parties controls intangible assets for which an arm’s length return is not easily determined. Since TNMM is applied to the party performing routine manufacturing, distribution or other functions that do not involve control over such intangible assets, it allows the appropriate return to the party controlling unique or difficult-to- value intangible assets to be determined indirectly.

TNMM may also be appropriate for use in certain situations in which data limitations on uncontrolled transactions make it more reliable than traditional methods. TNMM may be more attractive if the data on gross margins are less reliable due to accounting differences (i.e. differences in the treatment of certain costs as cost of goods sold or operating expenses) between the tested party and the comparable companies for which no adjustments can be made as it is impossible to identify the specific costs for which adjustments are needed. In such a case, it may be more appropriate to use TNMM to analyse net margins, a more consistent measured profit level indicator than gross margins in case of accounting differences.

10. In view of the above discussions, as also bearing in mind entirety of the case, we are of the considered view that the authorities below indeed erred in not applying the TNMM for ascertaining the arm’s length price of assessee’s transactions with the associated enterprises. We direct the AO/TPO to compute the ALP on the basis of the transactional net margin method. With these directions, we remit the matter to the file to the assessment stage for fresh determination of arm’s length price. As the matter is being remitted to the assessment stage, it will be open to the assessee to take such other plea, on merits, on ascertainment of ALP under the TNMM as the assessee may deem fit and the same will have to be disposed of by way of a speaking order, in accordance with the law and after giving a fair and reasonable opportunity of hearing to the assessee. All those issues regarding application of TNMM method, as on now, are academic and wholly hypothetical. With these directions, the matter is remitted to the file of the Assessing Officer.

11. In the result, the appeals are allowed for statistical purposes in the terms indicated above.