JD v Comptroller of Income Tax
[2005] SGHC 92

11 May 2005

Belinda Ang Saw Ean J:

1          In its grounds of decision ([2004] SGDC 245) dated 30 April 2004, the Income Tax Board of Review (“the Board”) affirmed the decision of the Comptroller of Income Tax (“the Comptroller”) in disallowing the claim of JD to deduct interest expenses in computing the amount of its income on which it was assessable to tax for certain tax years. This appeal by JD, the taxpayer company, which concerned only the disallowance of the interest expenses, involved the construction of s 14(1)(a) read with s 10(1)(d) of the Income Tax Act (Cap 134, 2004 Rev Ed) (“the Act”). The facts were all agreed to by the parties.

2          JD was (and still is) a public-limited investment holding company. At the relevant time, it received as its only income, dividends from shares held in other companies such as:

(a)        B, a wholly-owned subsidiary engaged in packing and trading edible oil products;

(b)        C, a wholly-owned investment company incorporated in 1981;

(c)        D, a wholly-owned subsidiary incorporated in 1982 and trading in consumer goods;

(d)        E, a wholly-owned subsidiary, engaged in piling and building construction, civil and structural engineering, renovation and retrofitting business and investments;

(e)        F, a public-listed company, engaged mainly in general insurance;

(f)         G, engaged in the business of warehousing and investment holding;

(g)        H, a wholly-owned property company incorporated in 1987 in Malaysia; and

(h)        J, a general trading company.

3          JD’s purchase of the shares in the companies in question were financed by means of (a) overdrafts and loans from banks and related companies at varying rates of interest and (b) JD issuing its own shares or obtaining interest-free loans from related companies. These funds were mixed and placed in its account with a Bank and from which bank account the collective funds were utilised not only towards acquiring JD’s share investments but also for its re-financing of earlier loans as well as advances to related companies.

4          The years of assessment in dispute spanned over a period of 12 years, from 1985 to 1996. Not all of the shareholdings in the companies in question declared dividends in the years of assessment in dispute. For instance, K, B and D did not declare dividends for any of the years of assessment in dispute. Some companies, on the other hand, did not produce dividend income for the following years of assessment:

5          It is convenient to reproduce, at the outset, the particular sections relevant to the appeal. For ease of reference, I have used the 2004 revised edition of the Act as the text of the sections under review is the same as the previous editions in use during the years of assessment in dispute. 

Charge of income tax

10.—(1) Income tax shall, subject to the provisions of this Act, be payable at the rate or rates specified hereinafter for each year of assessment upon the income of any person accruing in or derived from Singapore or received in Singapore from outside Singapore in respect of —

(a)        gains or profits from any trade, business, profession or vocation, for whatever period of time such trade, business, profession or vocation may have been carried on or exercised;

(b)        gains or profits from any employment;

(c)        (Deleted by Act 29/65);

(d)        dividends, interest or discounts;

(e)        any pension, charge or annuity;

(f)        rents, royalties, premiums and any other profits arising from property; and

(g)        any gains or profits of an income nature not falling within any of the preceding paragraphs.

Deductions allowed

14.—(1) For the purpose of ascertaining the income of any person for any period from any source chargeable with tax under this Act (referred to in this Part as the income), there shall be deducted all outgoings and expenses wholly and exclusively incurred during that period by that person in the production of the income, including —

(a)        except as provided in this section, any sum payable by way of interest upon any money borrowed by that person where the Comptroller is satisfied that the interest was payable on capital employed in acquiring the income; …

6          In so far as is material to this appeal, the dispute between the parties concerned the deductibility of interest described as expenses incurred in the production of dividend income in computing the amount of JD’s income on which it was assessable to tax for certain tax years. The taxpayer company pressed the point that investing in shares was employing capital to earn dividends. It incurred interest costs on the capital invested. So interest expended on capital employed to acquire the whole basket of investments in shares (which constituted one source of income) for the time being held by the taxpayer company was the cost of earning those dividends. Between 1985 and 1996, the taxpayer company earned dividends from its investments in shares while incurring interest costs on capital invested. Its dividends for each tax year ought to be assessable as a whole. JD’s net income in respect of the dividends was the difference between the gross dividends earned each year as a whole and the interest expense.

7          On the other hand, the Comptroller allowed interest expense as a deduction for only those shareholdings that produced dividend income. This was because the Comptroller regarded each shareholding as a separate investment and dividend income was separately assessable. The Comptroller thereafter aggregated the allowable interest expenses in respect of all the share investment counters to determine the total amount of deductible expenses for the particular year of assessment in determining the chargeable income for that year of assessment.

8          As a consequence of the difference in interpretation of the law (including its application to the share investments as passive income chargeable to tax), the aggregate chargeable income of JD for the years of assessment in dispute computed by the Comptroller was $83,484,337 compared to the sum of $74,694,762 as computed by JD. On the revised figure before the Board, the amount of tax in dispute was $2,497,841.74.

9          The Board, in accepting the Comptroller’s argument, clarified that the word “source” in s 14(1)(a) was not defined in the Act; it did not have a technical meaning and it should be interpreted according to its natural or ordinary meaning in the context of the provision. Given its ordinary meaning, the word “source” in s 14(1) meant “channel or stream of income” and the dividends from each of the different share counters constituted a separate and distinct source of income for the purposes of deductibility. Therefore, only interest expense incurred in producing the income from a particular share counter could be deducted from the dividend income generated by that shareholding. The Board at [38] of its grounds of decision explained:

Turning to the facts of the present case, each of the shareholdings in the subsidiary companies that are owned or held by the Appellant are distinct and identifiable. It was not in dispute that the funds used and the interest expenses in acquiring these separate shareholdings are identifiable. Applying the reasoning stated earlier on “source of income”, dividend income from one block of shares may be treated differently from that from another counter of shares where that block of shares is sufficiently distinct. Dividend income from each of the shareholdings in different companies would be regarded as a different source of income for the purpose of section 14(1)(a) ITA. This is consistent with the principle that section 14(1)(a) requires a direct nexus between expenses incurred and the source of income produced. It would fly against the reason of logic and the legal principle in Section 14(1)(a) that interest incurred for acquisition of shares in one company can also be deducted as expenses against the dividend income from the shareholding in a totally unrelated different company in which different interest expense was incurred using different borrowed money. Hence, only the interest expense on a particular counter of shares is deductible against income from that same counter of shares. Any excess of interest expense over dividend income from the shares is not allowable against other dividend income from other counters of shares.

10        The Board continued at [39] and [40]:

Based on the Statement of Fact, it [was] agreed between the parties that not all interest expenses were incurred on capital employed in acquiring income within the meaning of section 14(1)(a). …

On this basis, applying the reasoning in Andermatt’s case, we agree that using the words of section 14(1)(a) the Comptroller, in this case, cannot “… be satisfied that the interest was payable on capital employed in acquiring [the] income” because the loan (capital) to acquire the separate blocks of shares in these different companies did not produce dividend income in the relevant years.

11        The Board further ruled at [48] that the Total Assets Formula (for allowing an apportionment of interest into the non-income producing and income-producing elements in determining the amount of expenses deductible) adopted by the Comptroller was legally tenable and reasonable and the formula had been properly applied to the facts of the case.

12        Counsel for JD, Mr Leon Kwong Wing, submitted that the Board was wrong in concluding that the Act did not define the sources of income chargeable with tax. According to Mr Leon, sources of income chargeable with tax were clearly defined in s 10(1) of the Act. He cited Andermatt Investments Pte Ltd v Comptroller of Income Tax [1995] 3 SLR 451 (“Andermatt Investments”) where the Court of Appeal stated at 457, [16] that “[s]ection 10(1) sets out the various sources of income which are subject to tax”. He submitted that s 10(1) stipulated that the income chargeable with tax in this appeal was “dividends”. Mr Leon stressed the point that tax was charged on income from dividends as opposed to dividends from each investment.

13        Mr Leon argued that the Comptroller’s erroneous construction of s 14 of the Act then led to the adoption of the Total Assets Formula, which was a formula not provided for in the Act. It was a formula concocted to ameliorate the effect of s 14(1)(a) in purported exercise of administrative discretion conferred by s 5 of the Act. This approach, he argued, could not be right in the context of a fiscal legislation and was a “dead giveaway” that the Comptroller and then the Board had begun on the wrong premise in interpreting the statutory provisions. The Comptroller was under the misapprehension that the statute required the taxpayer company to identify to the Comptroller’s satisfaction an incoming of dividend against an outgoing of interest. He canvassed the view that the legislative material and cases from other jurisdictions cited by him indicated that dividend income was a single assessable item. It followed that against it the interest expense on the capital employed as a whole was deductible.

14        Responding, Mr Liu Hern Kuan for the Comptroller rejected as untenable JD’s contention that s 10(1) of the Act defined the classes of income which could be regarded as sources of income by themselves. Section 10(1) did not expressly state that the forms of income in s 10(1) were sources of income in themselves. JD’s reliance on one particular statement in Andermatt Investments was misplaced and taken out of context. There was no pronouncement by the appellate court that sub-ss (a), (b) and (d) to (g) of s 10(1) each constituted a single source of income for the purposes of deductibility under s 14(1). On the facts of that case, the Court of Appeal had to contend with one homogeneous block of shares of one company. The Board found that the High Court and Court of Appeal in Andermatt Investments supported the view that the word “source” in s 14(1) referred to a distinct and separate origin of income and it was not defined by the items of income expressly enumerated in s 10(1).

15        Mr Liu explained that the Comptroller’s position had always been that for interest expense to be deductible, the latter must be satisfied that the “interest was payable on capital employed in acquiring the income”. Under s 14(1)(a), income had to be produced and earned by the relevant expense before it was allowed as a deductible. On the authority of Andermatt Investments, there should be a close nexus between the income earned and the interest expense. The Comptroller had to be satisfied that the nexus existed before a deduction was allowed. Therefore, if the taxpayer company could not prove the linkage, the deduction would be disallowed since the onus of proof of the excessiveness of the tax was on the taxpayer: see s 80(4) of the Act.

16        As a matter of general practice, where a taxpayer acquired income-producing and non-income producing assets and incurred interest expense but was unable to relate the interest expense to the asset concerned, the Comptroller would adopt the Total Assets Formula as a condition of the Comptroller’s satisfaction. This Total Assets Formula attributes various amounts of interest to the various investments based on their proportion of cost in relation to the total assets. The Total Assets Formula that both sides had agreed to was         A   X  I

                                             C

where :

A (JD’s Basis)                       =       cost of all share investments financed by interest-bearing funds for each year of assessment in dispute

A (Comptroller’s Basis)         =       cost of income-producing share investments financed by interest-bearing funds for each year of assessment in dispute

C         =    total cost of assets as at the balance sheet date which were financed by interest-bearing funds for each year of assessment in dispute

I           =    the total interest expenses incurred by the taxpayer company for each year of assessment in dispute

17        It is clear, from this brief narration of the principal arguments of each of the parties, where the battle lines were drawn. On the facts as agreed by the parties, the whole of the interest expenses was not incurred in producing the dividend income. Unless the taxpayer company could maintain its interpretation of the statutory provisions under review, it must fail in its claim to deduct the whole of its interest expenses.

18        The system of taxing income, which includes dividends, begins with Part III of the Act, entitled Imposition of Income Tax. Section 10(1) provides that “[i]ncome tax shall ... be payable at the rate or rates specified … for each year of assessment upon the income of any person” derived from different kinds of gains or rewards of a revenue nature expressly enumerated in sub-ss (a), (b) and (d) to (g). They are treated as income of a person for the purposes of the Act. Dividend is, in my view, simply a description of a kind of income which s 10(1) enacts as the income (as contrasted with the capital) of any person. But that is not all, for the words “derived” or “accruing” in the subsection should not be forgotten or downplayed. Taken in context, the taxpayer’s activity or activities and property are looked upon as the originating cause of or from which the different type of gains or rewards enumerated and classified as “income” is derived, be it in or outside Singapore. In this case, the nature of the taxable income was dividend income arising from various shareholdings.

19        This interpretation is compatible with the reasoning of the High Court of Allahabad in Seth Shiv Prasad v Commissioner of Income-Tax, UP [1972] 84 ITR 15 at 18 (“Seth Prasad”) where Pathak J said:

[W]e find that income from dividends has been treated as one kind of income. What is the source of dividend income? It is the shareholding held by the assessee; and there can be as many sources of income as there are shareholdings. Shareholdings in different companies constitute different sources of income. The shares of one company may be treated by the assessee as a single shareholding. The assessee may also, for good reason, treat the shares of the same company as constituting a number of separate and distinct shareholdings. The shares may be divided into groups defined by reference to the circumstances in which they were acquired, or to the purpose for which they were purchased, that is, some as an investment holding and others as a share-dealer’s stock-in-trade, or to the category or class to which they belong, for example, whether they are preference or equity. There may be other criteria reasonably defining them into separate and distinct shareholdings, and, therefore, as distinct and separate sources of income.

20        Seth Prasad was cited by Mr Liu for the court’s observations on “what constitutes a source of income”. It is clear from the dicta that, when it comes to ascertaining a source of income, it is entirely a factual matter. L P Thean J (as he then was) in CH Pte Ltd v Comptroller of Income Tax (1988) 1 MSTC 7,022 adopted the dictum of Isaacs J in Nathan v The Federal Commissioner of Taxation (1918) 25 CLR 183 at 189:

The Legislature in using the word “source” meant, not a legal concept, but something which a practical man would regard as a real source of income. … [T]he ascertainment of the actual source of a given income is a practical, hard matter of fact. 

21        It is not the funds invested by the taxpayer company that the Act seeks to tax in s 10(1) but the dividend income that arose from activities of the taxpayer and its property (ie, the shares). To this extent, I am in agreement with the Board’s interpretation of the plain and ordinary meaning of the word “source” in s 14(1). As to the source of the dividend income, in my judgment, it must necessarily stem or originate from the particular share counter that yielded the revenue.

22        The approach is the same when it comes to dealing with deductions and other matters bearing upon the nature of taxable income under s 14(1)(a). The words “source chargeable with tax” in s 14(1) have reference back to s 10(1)(d). Section 14(1)(a) plainly limits the deductibility of interest expenses to those cases where the conditions specified in the subsection are satisfied. Section 14(1) uses the words “outgoings and expenses wholly and exclusively for the production of the income”. At the same time, sub-s (a) uses corresponding words such as “interest was payable on capital employed in acquiring the income”. The interest expenses are specific to “the income” and not “any income” chargeable with tax. The requirement for capital to be productively employed is evident. Lim Teong Qwee JC, who heard the appeal to the High Court from the decision of the Board, in Andermatt Investments Pte Ltd v Comptroller of Income Tax [1995] 1 SLR 66, said at 72, [18] of his judgment:

[T]he deductibility [of interest expense] is subject to a further condition. The Comptroller has to be satisfied that the interest was payable on capital employed in acquiring the income. While I accept that for the interest to be deductible from the rent income some relationship between the rent income and the money borrowed may be necessary I do not agree that any relationship, however substantial, is necessarily sufficient. There has to be a particular kind of relationship. The link is that provided in s 14(1)(a). To be deductible the interest must be payable on capital employed in acquiring and in acquiring the income and not just income or any income. [emphasis in original]

23        Chao Hick Tin JA, delivering the judgment of the Court of Appeal in Andermatt Investments, in construing s 14(1)(a), held that for there to be deductibility of interest expenses there had to be a direct link between the money borrowed and the income produced. In that case, the overdraft facility was used to purchase the shares in Wan Holdings Pte Ltd which owned the Hillview property at 1 Jalan Remaja. The taxpayer was not seeking to deduct interest expenses against dividend income, for by the time the deductions were sought, Wan Holdings Pte Ltd was already liquidated. There were no more shares and dividends to speak of. What the taxpayer wanted was to deduct interest expenses against rental income from the Hillview property. The appellate court agreed that the overdraft could not be attributable to the acquisition of rental income from the Hillview property which was actually transferred to the taxpayer by the liquidator as a return of capital in specie. In that case, the appellate court was concerned with the question of how the capital (the overdraft) was employed when it was first raised. Chao JA said that it was only when the borrowings were used to purchase the Hillview property that it could be said that the overdraft was employed in acquiring the rental income and the interest incurred could then be deducted from rental income.

24        It should also be noted that the main part of s 14(1) does not say that the expenditure incurred in the production of dividend was “for any period”, whether it is for the current period, an earlier period or a later period in respect of which tax is chargeable. Instead, it provides that the deductibility of a particular expense like interest expense in a given basis period must depend upon it having been incurred in the same period in the production of dividend income that is chargeable with tax. Section 14(1) talks of “outgoings and expenses incurred” and by sub-s (a) expenses include “interest” payable in the same relevant taxation period as the dividend income chargeable with tax in the hands of the recipient.

25        Mr Liu Hern Kuan took issue with Mr Leon’s submission that “as a general proposition, it seems … incontrovertible that any coherent system of yearly income taxation must allow costs incurred to earn income to be deducted in the year the costs are incurred, whether the income they are aimed at produced is earned in the immediate tax year or in the future”. In the context of s 10(1)(d), Mr Liu was right when he said that Mr Leon’s submissions contradicted s 14(1) which required the interest expense to be incurred in the same period in which income was produced.

26        In a s 10(1)(d) case, if there was no dividend income or if dividend income was less than interest expense, no carrying forward of losses would be allowed under s 37 of the Act. Consequently, any interest expense incurred in a year in which that investment did not produce income would be lost as it cannot be carried forward. It follows that if no dividend income was produced by a particular share investment, interest expense may not be set off against another share investment. I would illustrate this using rental income, which is chargeable with tax under s 10(1)(f) and like dividend income is considered passive income. An investment holding company X owns two houses for rental. Interest expenses are being incurred from bank borrowings to purchase both houses. One house is rented but the other is vacant, as X has not found a tenant. The interest expense incurred and attributable to the vacant house cannot be used to offset rental income for the other rented house under s 14(1)(a) even though it arises in the same year of assessment.

27        In Andermatt Investments there was no direct relationship between the money loaned, the purchase of the property and the income produced. A similar problem plagued JD. JD was not able to satisfy the direct nexus test. The shares so purchased had to be income producing before a deduction for interest expense was allowed and JD ran into difficulty with that. In the context of this case, I agreed with Mr Liu that under s 14(1)(a) income must be produced and earned by the relevant interest expense before it is deductible. Hence, the interest expenses attributable to the shares which did not produce income could not be set off against income from those shares that did not incur the interest expense but did produce dividend income.

28        It therefore rested on the taxpayer company to identify the interest expenses and show the amount of the expenditure which could be said to be referable to the production of the receipts (ie, dividend income) in the year of assessment. In the present case, different shares were acquired at different times using different loans (both interest-bearing and non-interest bearing funds). JD’s position was that all shares were a single source, dismissing any need for a direct nexus between the capital employed in acquiring the income and the interest expense. The Board disagreed and applied the direct nexus test. It said at [32] of its grounds of decision:

Hence, applying this test of a direct link, it necessarily follows that only the interest expense incurred for the overdraft facility obtained to acquire the particular shareholding can be deducted from the dividends from that same shareholding for the same period.

29        Mr Leon relied on Malaysian, Indian and Australian cases to support his submissions that dividends need not have been earned before interest on capital borrowed to acquire shares was within the ambit of the words “incurred solely for the purpose of making or earning such income, profits or gains” in s 12(2) of the Indian Income Tax Act. In short, the Board was wrong to think that the taxpayer’s cost in interest could be taken into account for income tax only if and when there was a return in the share investment. The Indian cases were Ormerods (India) Private Ltd v Commissioner of Income-tax, Bombay City [1959] 36 ITR 329 at 334 and 335; K Appa Rao v Commissioner of Income-tax, Madras [1962] 46 ITR 511 at 514; P V Mohamed Ghouse v Commissioner of Income-tax, Madras [1963] 49 ITR 127 at 132 and 135; and M N Ramaswamy Iyer v Commissioner of Income-tax, Kerala [1969] 71 ITR 218 at 225. Mr Leon referred me to Federal Commissioner of Taxation v Total Holdings (Australia) Pty Ltd (1979) 43 FLR 217 where the commissioner of taxation accepted that a holding company would necessarily incur interest costs and such interest was a permitted deduction because shares were by their nature inherently capable of generating dividends. This was so even though there was no guarantee that these investments would ever prove profitable. New Zealand tax practice appeared to be in the same vein: see Staples’ Guide to New Zealand Tax Practice (Brooker’s, 60th Ed, 2000) at paras 830.40 and 830.50.

30        The two Malaysian cases relied on by JD in the appeal were P Securities Sdn Bhd v Ketua Pengarah Jabatan Hasil dalam Negeri (1995) MSTC 2,256 and Ketua Pengarah Hasil dalam Negeri v Multi-Purpose Holdings Bhd [2002] 1 MLJ 22. Mr Leon maintained that the factual situation in these two cases was on all fours with the present appeal before me. Whilst that may appear to be so, the reasoning there was that to treat each counter of share investment as a separate source of dividend would be to “disintegrate the six grouping or categories of income” chargeable to Malaysian tax in s 4 of the Malaysian Income Tax Act 1967 and that would be contrary to that legislation.

31        Mr Liu submitted that Mr Leon’s reliance on the foreign tax cases was unhelpful in that differently-worded tax statutes were involved and one could not indiscriminately apply principles from foreign tax cases. The foreign cases and statutory provisions were different from s 14(1)(a). The Board agreed with the Comptroller, citing as guidance Rhodesia Metals Ltd v Commissioner of Taxes [1941] 9 ITR (Suppl) 45 where the Privy Council at 51 stated:

Their Lordships ... desire to point out that decisions on the words of one statute are seldom of value in deciding on different words in another statute; and that different business operations may give rise to different taxing results.

32        To illustrate, s 12(2) of the Indian Income-Tax Act 1922 is differently worded. Our s 14(1) makes no reference to “purpose”, although there is reference to expenses wholly and exclusively incurred. Another difference is the word “paid” which refers to the same period the income was produced and expenses incurred. Section 12(2) of the Indian Act has no restriction on “period”.

33        Before the Board, JD cited the same Malaysian cases. The differences in both the tax statutes were identified and fully discussed by the Board. This was what the Board said at [23] to [26] and [28]:

… Firstly, the Malaysian statute on deductibility for interest expense is different from s 14(1)(a) of our Income Tax Act. The three Malaysian cases had relied on the Malaysian deduction provisions. Our section 14(1)(a) requires the Comptroller to be “satisfied that the interest was payable on capital employed in acquiring the income.” The Malaysian provision, section 33(1) of the Malaysian Income Tax Act, does not have such a requirement. The term, “… where the Comptroller is satisfied…” authorises the Comptroller to apportion the interest expense for deduction purposes, in a manner which is acceptable to him. While it does not confer unfettered discretion on him to apportion as he wishes, the apportionment it is submitted, ought to be made in accordance with basic tax principles, and in particular, on the basis that the interest ought to be incurred wholly and exclusively in the production of income.

An additional significant difference in Malaysian legislation is the presence of section 5(1)(b) of the Malaysian Act. The Multi-Purpose case had based its position on this provision. … In contrast, our Act does not have an equivalent of section 5 of the Malaysian Income Tax Act which contains the words “source consisting of a business”. …

Secondly, the word “source” ... appears many times in the Malaysian Income Tax Act … The source concept is therefore extensively mentioned in the Malaysian Act, which is not the case for the Singapore Act. Some provisions for example, sections 11, 31, 54, 60, 60E and 60G deem certain income-producing activities to be a separate source of income. It would appear that the Malaysian legislation attempts to fix the source of many forms of income. …

Thirdly, P Securities … and the Multi-Purpose case should not be followed because these cases applied the two UK cases of [Diggines v Forestal Land, Timber & Railways Co Ltd 15 TC 630] and [Merrifield v The Wallpaper Manufacturers Ltd 16 TC 40] wholesale. As these two cases were based on the particular provisions of the UK Income Tax provisions, their relevance to the Singapore context is limited. This is because the structure of the UK Income Tax Act is vastly different from the Singapore Act. The UK Income Tax Act is divided into many ‘Schedules’ and these have been held as mutually exclusive. …

On the other hand, there is only one charging provision under the Singapore Income Tax Act s 10(1), while in the UK, there are many charging provisions, each creating separate sources of income. … Under the UK Act, there would not be any further justification for further sub-division or ‘disintegration’ of income into further sub-categories. …

34        Obviously, the concern is in simply accepting or adopting interpretations which have been placed upon particular words in other jurisdictions without according due weight to the particular statutory framework of that jurisdiction. Regard may sometimes be had to authorities in other jurisdictions if it is possible to balance the differences within the statutory schemes. That was not the case here. Not only was the wording of the provisions under comparison different, the particular statutory scheme under review was different from the other jurisdictions referred to by Mr Leon. Whilst the Malaysian tax system might have once been the same, it was no longer the case as our taxation statute had undergone substantial amendments. Reliance on the 1947 Income Tax Ordinances of Malaysia and Singapore was not helpful. As counsel for the Comptroller pointed out, the two statutes are now quite differently worded. Malaysian provisions would not be applicable where concepts are differently developed in Malaysia. More to the point, after considering the submissions made on appeal, I agreed with the Board that the material and cases from the various jurisdictions cited by Mr Leon should not be followed.  

35        As stated, it rested on the taxpayer company to quantify the interest expenses where dissection was necessary to yield a logical or approximate result to show the amount of the expenditure which could be said to be directly referable to the production of the receipts of dividend income. The resources to invest were drawn from a mixed pool of funds. While the aggregate amount of interest payable on the aggregate capital invested for any period was known, matching every item of income to an item of expense was difficult. It must be remembered that deductibility in s 14(1)(a) is subject to the Comptroller being satisfied that the interest was payable on the capital employed in acquiring the income. In this case, the Comptroller was not so satisfied that all interest expenses claimed by JD were payable on capital employed in acquiring income. Rather than disallowing the entire amount of interest expense, the Comptroller, in exercise of its administrative discretion conferred by s 5 of the Act, applied the Total Assets Formula. I would not consider the Comptroller’s application of the Total Assets Formula as an imposition on the taxpayer company of an extra-legal apportionment formula merely because he expressed himself, in exercise of administrative discretion, not to be satisfied with the taxpayer’s proof that the interest was payable on capital which earned the income. There was no stated criterion as to how that discretion should have been exercised by the Comptroller. It could, of course, be challenged on established grounds such as unreasonableness by the court. So long as it was reasonable and not arbitrary, imprecise or elusive, the quantum of deduction was a matter for the Comptroller to decide.

36        Mr Liu was quick to point out that counsel for the taxpayer company was inviting the court to examine the legality of the Total Assets Formula when JD had earlier agreed and accepted before the Board that the formula ought to be applied. He referred me to para 9 of the Agreed Statement of Facts:

Both [JD] and the Comptroller agree to the application of the Total Assets Formula (hereinafter “the Formula”) in determining the amount of allowable interest expense. However, they differ as to the method of application of the Formula, due to the reasons set out in paragraphs 10 to 12. The parties have since agreed to have the appeal argued on the basis of the Formula, as applied by both parties.

37        JD should not be permitted to do an about-face. I was satisfied that the Comptroller’s basis for numerator A in the Total Assets Formula was not analogous to extra-statutory concession but was a sensible attempt by the Comptroller to explain how s 14(1)(a) might be applied according to Andermatt Investments. As a result, the Comptroller applied the Total Assets Formula in a different way from JD. Besides, the apportionment formula is not unique to Singapore, having been used in Hong Kong and South Africa. The Board found the Total Assets Formula adopted by the Comptroller to be legally tenable and reasonable and to have been properly applied to the facts of this case. I saw no reason to disturb the Board’s findings.

38        Having decided the principal point of the appeal, which was on a construction of s 14(1)(a) read with s 10(1)(d), in favour of the Comptroller, there was no necessity to deal with the argument under s 15(1)(c) of the Act. Separately, there was another reason. Mr Liu raised s 15(1)(c) as an additional ground to s 14(1)(a). In my view, it was too late for the Comptroller to raise s 15(1)(c), having applied in this case the Total Assets Formula. For completeness, I should mention that in deciding on the principal point of the appeal, it was also not necessary to deal with the myriad arguments and cases relied upon by both sides.

39        In the result, I dismissed the appeal with costs. 

Appeal dismissed.