GRANT THORNTON KESSEL FEINSTEIN
1. The desirability of the tax

I fully realise that Parliament has the power to vote against this Bill but that, as a practical matter, this is extremely unlikely. Nevertheless, a comment on the Bill cannot properly be made without commenting on the tax itself

It has been said that when a government introduces complex new tax legislation, it is a boon to private sector tax consultants such as myself as it gives us new opportunities to charge fees to our clients. This, of course, it absolutely true. However, in a communication such as this I am writing not from the view point of a private sector tax consultant, but rather from the view point of a concerned South African citizen seeking what is best for the country as a whole.

Consequently, regardless of the potential damage that it may do to my back pocket and those of my partners, I would strongly urge the Committee to recommend to Parliament not to pass this legislation, for the simple reason that I do not believe that such a tax is the right thing for this country at this (or, for that matter, any other) time.

It is true that most of our trading partners have this tax in their armoury of fiscal legislation. But it is also true, that while we are introducing the tax, a number of them are easing the rules significantly, either my lowering the rates or granting further concessions, and from what I read, they are noticing the benefits to their economies as a result. Why are we, in the face of this evidence, seeking to introduce a tax which is known to have an adverse effect on markets and the investment environment?

Of course, it is not always wise to compare a country such as South Africa with first-world countries such as the UK, USA, or Australia. Sometimes we should take note of other emerging markets, such as Kenya, which introduced capital gains tax and then abolished it 15 years ago.

While acknowledging that the tax, viewed from a micro point of view, is not a good tax in the sense that the revenue it generates is not justified by the amount of resources that have to be invested in order to collect the tax, nevertheless the Ministry of Finance has justified the tax on the basis that it is an effective anti-avoidance measure as regards income tax. The theory here is that it is possible to restructure transactions in such a way that one forgoes income, which is taxable, and instead one receives a tax-free capital gain. Thus, by introducing a capital gains tax, one counters this tendency.

This logic is absolutely correct in theory, but, because of practical elements, there is a weakness in the argument that needs to be brought to your attention. Having capital gains tax to plug a hole where income tax is avoided will only have the desired result where the effective rate of capital gains tax is the same as the rate of income tax that would otherwise have been payable. But given that, apart from the other exemptions and exclusions, only 25% or 50% of the gain is subject to the tax, the incentive to "convert" income into capital remains as strong as ever. Thus, only if the capital gains tax rate was 42% for individuals and trusts and 30% for companies would the argument truly hold. And in no country in the world of which I am aware and against which we benchmark ourselves is the effective rate applicable to capital gains tax ever equal to the nominal rate applicable to income tax, simply because there are always allowances, indexation relief, lower rates, and the like.

It has also been stated that capital gains tax introduces certainty. This is, with respect, incorrect. As long as there is the differential in the effective rate of tax the debate as to whether a gain is on capital account or revenue account will continue unabated. Certainty can only be introduced legislatively. Section 9B of the Income Tax Act deems profits on sales of listed shares held for more than five years to be capital in certain circumstances. This would have to be extended to all assets and the period reduced significantly, say to one or two years.

Thus, given the empirical and undisputed evidence that capital gains tax is an inefficient tax and does not of itself justify the resources invested for its collection, and given further the empirical evidence that it does affect markets, the economy and liquidity adversely, and given further the weakness relating to tax avoidance, I really cannot see a place for this tax in this country.

I am, however, a pragmatist. As I said, I realise that, while it is legally possible that Parliament will refuse to agree to pass this legislation, from a practical point of view this is extremely unlikely. Accordingly, on the assumption that the legislation will, indeed, be introduced, I deal below with other issues actually related to the legislation itself.

Timing of implementation
The SARS was given a Herculean task after the last Budget to bring in major new legislation affecting the taxation of foreign dividends, the move to a worldwide tax basis and the introduction of capital gains tax. The SARS is to be commended for the amount of effort they have put in to rise to this task.

However, with respect, I believe that it is too much of a burden to have placed not only on the SARS but on the country as a whole to deal with so much change in such a short period of time. Indeed, one merely has to observe that the final version of this legislation is likely to be made public and passed by Parliament a mere three to four weeks before its implementation date. With respect, this is unreasonable.

Apart from anything else, it is unreasonable to expect taxpayers to adjust their accounting systems and records in such a short period of time to enable them to adapt to the new legislation and to rearrange their affairs so as to operate efficiently within its parameters. For example, there are some fairly onerous reporting requirements placed on unit trust and asset management companies in terms of the new sections 70A and 70B of the Income Tax Act, and much of this information needs to be captured commencing 1 April 2001. It is unreasonable to expect accounting and computer systems to be suitably modified by that date.

I would therefore wish to recommend that if this legislation is to be introduced, its implementation should be delayed. Ideally this should be by one year but if it is absolutely necessary that it be introduced in the current fiscal year, a delay of, say, six months would not be an unreasonable request, nor one which would result in any loss of face by any person.

To be effective from 1 April 2001 the legislation must be passed in March at the latest. The Committee will cease its hearings on 16 February 2001. There have no doubt been voluminous submissions to this Committee and to the SARS in addition to these oral submissions. Can all the submissions be properly studied and debated and the legislation still redrafted in the few weeks remaining before the Bill is tabled in Parliament such as honestly to do full justice to the submissions?

The fact is that income tax was first introduced into South Africa in 1914. We have managed to live without this tax for the last 87 years. Surely we can survive another six months or year without it without the sky falling in!

The threat of economic double taxation
In a corporate environment there is a strong possibility, and in some case inevitability, that the same gain will be subject to tax more than once from an economic point of view, albeit it not legally. For example, assume a chain of companies comprising company A which owns all the shares in company B which owns all the shares in company C. The lastmentioned company sells its business and makes a profit on sale of goodwill. That profit will be subject to capital gains tax. Company C is then wound up and the profit distributed to company B. This will be a profit in B’s hands on disposal of the shares and will, again, be subject to capital gains tax. When company B is wound up and the profit distributed to company A, tax is payable yet a third time and if company A is ever wound up and distributes the profits to its shareholder, Mr or Mrs X, the tax is payable for a fourth time. In this case the effective rate is 54%! Even in a simple situation of Mr or Mrs X owing shares in a company or CC, if the business is sold (as opposed to the shares or members’ interest) the tax will be payable at least twice, and the effective rate is 25%, ie. 15% paid by the company/CC and 10,5% on 85% paid by Mr and Mrs X.

I would urge that provisions be written into the legislation to prevent this cascading of taxes, in a similar way in which a cascading of secondary tax on companies (STC) is prevented as a dividend passes up through a chain of companies. In the example given above, if a dividend were declared by company C it would pay STC but no further STC would be payable as that dividend passes up through the chain of companies until it eventually is paid to the shareholders of company A. In such case the effective rate is still 24.4%.

Assuming such an exemption is not granted, the problem of such a cascade of tax is very real in South Africa as many of the groups have deep, as opposed to flat, structures and pyramid arrangements. Such structures and arrangements were not disadvantageous from a tax point of view and, indeed, held other commercial advantages. If the exemption is not granted, it is necessary for groups of companies to rearrange their affairs so as to flatten their structures in order to minimise this economic double taxation.

To some extent the draft legislation caters for this but, with respect, it does not nearly go far enough. Clause 24 of the Bill amends section 39 of the Taxation Laws Amendment Act, 1994 to recognise that in an approved group rationalisation any movement of assets within the group will not give rise to any capital gains tax. However, there are, in my view, two major deficiencies in the concession:

In order to qualify for a rationalisation under section 39 the sole or main purpose of the rationalisation must be essentially to obtain commercial advantage and not for the purpose of tax avoidance. Thus if one had to seek to rationalise a group for the sole or main purpose of flattening the structure to avoid the adverse results described above, no approval could be given to such an application because tax avoidance is the sole or main purpose. I strongly believe that the dispensation contained in section 39 should be available even if the sole or main purpose of the rationalisation is to make the structure more efficient as a result of the introduction of capital gains tax. If it is felt that this could lead to abuse, one can control this by, for example, giving a restricted period during which such applications will be entertained, eg. within one year of implementation date.

Even if the sole or main purpose is not tax avoidance, but a significant benefit of the rationalisation will be the avoidance of a cascade of taxes, section 39 is only available where the holding company is listed on the Stock Exchange or it has a fixed capital of more than R75 million. As part of the Bill, this cap of R75 million is reduced to R50 million. While the reduction is welcomed, nevertheless even at R50 million a huge proportion of private companies is excluded from the legislation and, hence, from taking any steps which their larger counterparts could take in order to avoid the unforeseen consequences of a new tax. This is discriminatory against smalI and medium enterprises (SMEs). I strongly suggest that access to section 39 be made universal. The SARS have indicated in the past that the reason for the cap is that they do not have the resources to deal with a flood of applications. This argument now rings hollow in light of their insistence that they are now adequately resourced to handle worldwide tax and CGT.

Applicability to non-residents
In general terms, no capital gains tax is payable by a non-resident except in the case of –
immovable property, and
assets relating to a permanent establishment, branch or agency of the non-resident in South Africa.

I have no quarrel with 4.1.2 but I think that the taxation of immovable property referred to in 4.1.1. should be seriously reconsidered.

There is no standard approach among the countries as to what is included or excluded in the case of non-residents. Some countries adopt the approach we have taken, namely, to exempt non-residents from the tax except in the case of fixed property. Other countries such as Australia tax everything owned by a non-resident, while yet other countries, such as the UK, exempt everything.

Frankly, I can see the logic in exempting everything and I can also see the logic in taxing everything, but I cannot see the logic of taxing only immovable property and exempting everything else. What is so special about immovable property that it should be subject to the tax, but not any other asset, such as shares?

But if the tax is going to be imposed on non-residents, it must be enforced with vigour, because to introduce legislation which cannot, and will not, be enforced is to breed disrespect for the law. And I fear that there are inadequate mechanisms in the legislation, or in the SARS, to ensure that the tax is collected when a non-resident sells fixed property in South Africa.

Where the fixed property is a commercial property yielding rental, or one of the assets of a permanent establishment or branch, the non-resident will be on charge with the SARS and will be rendering annual income tax returns. If it sells the property it will have to record same in its tax return and the amount will be duly assessed and tax will be levied (and, one presumes, paid even if the non-resident no longer has any further dealings in South Africa). But where the property is not rental-producing, eg. a private game farm, holiday flat in Clifton, and the like, the SARS has no knowledge of the non-resident nor is the latter required to make his presence known to the SARS. When the property is sold the funds will simply be remitted abroad and no tax will be payable. What is more, in very many cases the non-resident will extract his or her money entirely innocently simply because they are not aware that there is capital gains tax to pay.

Some countries deal with this problem by imposing a withholding tax on proceeds of sale of fixed property by a non-resident in terms of which the South African purchaser is obliged to withhold the tax when paying the proceeds over. If the amount exceeds that which is properly payable, the non-resident would be free to submit an appropriate return to the SARS and receive the refund of the excess.

Given the amounts of revenue that are likely to be involved (and I freely admit that I have no knowledge of the extent of this revenue) I am inclined to the view that non-residents should be exempt from the tax on fixed property itself. But if it is imperative to include them, I think it is equally imperative to ensure that proper mechanisms exist for the collection of that tax and the policing of the system.

Residential accommodation
The original proposal was that residential accommodation should be completely exempt from the tax. The draft Bill now provides that the exempt portion will not exceed R1 million. While this might seem like a significant sum, and therefore will bring only within the net those extremely wealthy individuals in South Africa, I believe that that situation is more apparent than real. I fear that, economically speaking, a far greater number of people will be prejudiced than might otherwise be thought to be the case.

In the first place, unlike other assets, residential accommodation tends to be held for a lengthy period. While it is true that some people buy and sell houses within a few years, at the same time many people spend many years in their residences (I, for one, have been living in my house for more than 20 years). If one takes an average dwelling in a mid-market suburb, say in Johannesburg, one will find that the house is likely to cost in the region of R500 000 to R600 000. If one keeps that house for a period of 20 years, given movements in the market, the economy, inflation, and the like, it is easy to envisage that a profit of more than R1 million will be made when the house is sold. Given that, in the normal course, one tends to plough the proceeds of sale of one residence into the next residence, and given the rise in property prices generally, one can see that the individual will be disadvantaged because there will be less money for reinvestment in the new home.

An obvious answer to the point I raise is that it is likely that the R1 million exemption will not still be R1 million in 20 years time but will have increased as well. I am afraid that I would take this answer with a great deal of scepticism. The fact it that, historically, when a monetary cap is inserted in fiscal legislation, in the vast majority of cases the numbers are not adjusted to keep pace with inflation. To quote but 4 examples:

· · The Seventh Schedule to the Income Tax Act deals with the taxation of fringe benefits, and one such fringe benefit which is taxed is an interest-free loan. To avoid administrative overload, the legislation exempts from tax the benefit on a casual loan of not more than R3 000. This exemption was inserted into the legislation when the Seventh Schedule was first enacted in 1984, and has never been increased since then.

In section 11(u) deductible entertainment expenditure was capped at R2 500 in 1984 and has never increased since then.
The exempt portion of a gratuity payable on retirement or retrenchment (section 10(1)(x)) was last increased to R30 000, also in 1984.
In the Estate Duty Act, the exemption threshold was, ironically enough, also set at R1 million – thirteen years ago!

In fact, the vast majority of monetary limits in the Income Tax Act have never kept pace with inflation. (Some countries deal with this problem by automatically indexing any monetary amount contained in their legislation, and maybe this is something we should consider for South Africa as well.)

What is more, no account is taken of the economic reality of what a home truly costs. Paragraph 22(2)(d) of the Eighth Schedule excludes from the base cost of the asset any borrowing costs. This is entirely logical because the cost of an asset to a person who pays cash should be the same as the cost to a person who borrows to finance the acquisition of that asset. (That is not to say that interest should not be allowed – it should be, because the gain is being taxed and thus the borrowing cost should be allowed, but not as part of base cost.) However, in the case of a primary residence I think one should take a slightly different view. It is only a tiny minority of individuals who are in the fortunate position of being able to pay cash for their homes. The vast majority of people require mortgage bond finance. Thus while, legally speaking, it is clear that interest does not form part of the cost of an asset, from an economic point of view, in this narrow context, I would submit that borrowing costs should be included. After all, if one had to ask someone in ordinary conversation what his or her house costs, that person would, in the normal course, tend to look at the monthly instalment (inclusive of interest) as being the cost of acquiring the house.

It should also be remembered that, while interest may not form part of the base cost of a primary residence in other countries, eg. the UK or the USA, in those countries the income tax legislation allows a tax deduction for interest paid on a mortgage bond related to a primary residence. We offer no such facility in this country.

Accordingly, it is my strong suggestion that the cap of R1 million be removed and that the entire profit remain tax-free as per the original proposals. But if this is not considered acceptable, I suggest that the following amendments be made in order of preference:
The legislation be amended to –
· provide that the R1 million cap be automatically increased each year by the rate of inflation, and
despite the provisions of paragraph 22(2)(d), borrowing costs on a mortgage bond to finance the primary residence be allowed to increase the base cost.
The amount of R1 million be increased significantly, eg. to R5 million.

Donations to public benefit organisations
Paragraph 50 of the Eighth Schedule states that a capital gain arising from the donation of an asset to a public benefit organisation (PBO) will only be disregarded if the donation itself would qualify as a tax deduction under section 18A. I would like to recommend that capital gains be disregarded in respect of all such donations, irrespective of whether or not the donation would qualify as a tax deduction.

It is easy to see why the provision was introduced, because otherwise tax could be avoided. Thus a person wishing to make a sizable cash donation to a PBO might find that he or she has to realise certain assets in order to do so. Realising the asset will result in capital gains tax payable and thereafter the cash will be paid as a donation. To avoid this tax one could simply donate the asset to the PBO and let the latter sell it.

On the other hand, one could look at this in a different way. If by donating the asset, and letting the PBO dispose of it, one avoids capital gains tax, it will act as an incentive to wealthy people to make sizable donations to PBOs. Alternatively, if the proceeds on sale are, say, 100 and the capital gains tax is 5, the wealthy person might then donate only 95 in cash thereby effectively placing the tax burden on the PBO. If the asset is donated directly and the PBO sells it, it will realise 100 and therefore be better off. (It is interesting to note that in the USA if a donation in kind is made the donor receives a tax deduction equal to the market value of the asset whereas in South Africa the deduction is equal only to the cost of the asset.)

Rollovers
At present, the draft legislation does not exempt sales of assets within a group. The Explanatory Memorandum indicates that this will be examined further but that, in any event, there are unlikely to be any significant gains arising within the first few years of implementation.

This can lead to considerable difficulties. It is very common to transfer assets within a group at cost. This is no longer possible because sales between connected persons are deemed, in terms of paragraph 26 of the Eighth Schedule, to be at market value. On the other hand, realising profits on sales within a group is often undesirable and they have to be reversed when presenting consolidated financial statements to shareholders, causing significant difficulties in accounting systems.

I have also noticed that this provision is not limited for the purpose of capital gains tax but would appear to apply for all purposes under the Act. This will trigger tax recoupments if, say, manufacturing plant is transferred at book value within group companies but is deemed to be transferred at market value if the latter is higher. What is more, there are provisions in the legislation which limit the purchaser in a group situation to claim the depreciation on a value equal to the lower of cost to the seller and market value. Thus an intragroup transaction will trigger two different treatments within the same Act.

A further problem with this approach is that there is no certainty that the matter will ever be put forward for future debate. I would therefore prefer it if an alternative approach were to be taken: Introduce the legislation now and if, in years to come, it appears that it is unduly prejudicial to the fiscus, the legislation can always be amended or even repealed. But at least it should be put on the statute books now.

Additionally, unlike in overseas countries, eg. the UK, no rollover relief is given in respect of a share for share exchange. Thus if someone exchanges shares in company A for shares in company B, unlike in the UK tax will have to be payable in South Africa. This is extremely harsh. Normally when one sells an asset for cash one then has the cash available out of the proceeds to pay the tax. In the case of a share for share exchange (or any exchange of assets) there is no cash available to pay the tax and one has to find the money out of one’s other resources. I cannot, with respect, see why the SARS has set is face against this common, and logical, concession. There is precedent for deferring tax in the case of an exchange of assets. Section 24A of the Income Tax Act allows a property or share dealer to exchange the property or shares for shares in a listed company without paying tax, and tax is only payable when the listed shares are sold. Why can the same principle not be incorporated into the capital gains tax legislation?

Attribution of gains
Paragraphs 56 to 59 of the Eighth Schedule have the effect of causing a capital gain of one person or entity to be taxed in the hands of another. This applies where the latter has made a "donation, settlement or other disposition" in favour of the former. In principle I have no quarrel with this piece of anti-avoidance legislation. But I fear that, they way the legislation is drafted, it could result in undue prejudice which is not, I believe the intention.

Paragraphs 56 to 59 have their counterparts in respect of income tax in section 7 of the Act. In that case income accruing to someone by reason of a donation, settlement or other disposition made by another will result in the tax on that income being paid by the donor. The courts have held that if, for example, an individual makes an interest-free loan to a family trust, the non-charging of interest represents a donation which results in the section becoming applicable. Thus if the trust earns income, so much of that income as would have been earned by the lender had he or she charged a market-related rate of interest will be taxable in the lender’s hands as arising from a donation.

Assume that X makes an interest-free loan of R1 million to a family trust which then invests the funds and earns, say, R200 000. Assume that a market-related rate of interest is 15%. Section 7 of the Act will then apply so that the lender (donor) is taxed on R150 000 and the trust is taxed on R50 000 even though the lender physically earned nothing and the trust physically earned R200 000. If then the trust sells the asset and makes a capital profit, paragraph 58 of the Eighth Schedule would then purport to attribute part of the gain to the lender by virtue of the non-charging of interest. But if the lender has already been subject in full to income tax because of the non-charging of interest, how can that lender then again be subject to capital gains tax for the very self same reason? Either the trust earned the income because of not charging interest or it realised a gain, but it cannot be both.

It was my understanding that, when paragraphs 56 to 59 were to be introduced into the legislation, they were to apply only to the extent that section 7 had not already applied, but that both were not intended to apply to the same donation. However, there is nothing in the legislation that I can see that creates any sort of hierarchy or exclusion, and I believe that it is proper that the Eighth Schedule should attribute a capital gain only to the extent that an inadequate amount of income tax has been paid by virtue of section 7.

Estate duty
When someone dies there is a deemed disposal for capital gains tax purposes giving rise to capital gains tax payable by the estate. Estate duty is also payable by the estate on the net value of the assets included therein.

In recognition of the interplay between these two taxes, it is proposed that estate duty be reduced from its present level of 25% to some lower percentage which has not yet been made known. I believe that, unpopular as my suggestion may seem at first blush, the time has possibly come to repeal the estate duty legislation (and donations tax as well for the same reason).

An estate duty is essentially a social tax payable when assets of one person devolve upon others on the former’s death. As a revenue-raising mechanism, historically and internationally it is insignificant. One merely has to look at the amount of revenue raised in South Africa from estate duty to see how truly miniscule this amount is in relation to government revenues. However, being a social tax, it clearly has widespread appeal.

Given that capital gains tax will in any event now become payable on death (or by donation) the social element of the tax is present and one can therefore, with justification, state that the capital gains tax has supplanted the estate duty (and donations tax) as a social tax.
The experience in other countries is varied. The US has both capital gains tax and federal estate tax, the UK has both capital gains tax and inheritance tax but Australia and Canada have capital gains tax only and no form of estate duty or donations tax. It would therefore not be internationally out of step if South Africa were to abolish estate duty and donations tax once capital gains tax is introduced.

Obviously there is a difference on what the tax is payable. Capital gains tax is payable only on the excess of market value over cost at the date of donation or death whereas estate duty and donations tax is payable on the market value of the asset itself. Moreover, someone who dies, say, on 1 September 2001 would pay virtually no capital gains tax because the increase in value since 1 April is likely to be very small, but would continue to pay estate duty on the gross value.

To counter this problem, and bearing in mind the small amounts of tax raised by estate duty and donations tax, I would like to suggest that legislation be introduced now to abolish estate duty and donations tax but with effect from, say, two years after the implementation date of capital gains tax.

Finally, I would like to thank you and your committee for affording me this opportunity to make my views known. It is intended an honour and privilege to be here.

ERNEST MAZANSKY
National Chairman and Senior Tax Partner