Capital Gains Tax in South Africa
Briefing by the National Treasury’s Tax Policy Chief Directorate
to the Portfolio and Select Committees on Finance
Wednesday, 24 January 2001

1. Introduction
In the 2000 Budget, the Minister of Finance announced the introduction of a capital gains tax with effect from 1 April 2001 as part of a wider tax reform effort aimed at enhancing the integrity, efficiency and equity of the South African income tax regime. This seeks to address some of the existing avoidance opportunities such as recharacterising ordinary (taxable) revenue into capital revenue, thereby escaping the tax net. Hence, the introduction of a capital gains tax serves as a backstop to the income tax system, which enables the fiscus to broaden the tax base and ultimately reduce the rate structure, for a given revenue requirement.

The idea of taxing capital gains is not new in South Africa. In 1969, the Franzsen Commission proposed a limited form of capital gains tax on immovable property and marketable securities, while the Margo Commission in 1986 recommended that capital gains should not be taxed. Most recently, the Katz Commission considered the merits and demerits of a capital gains tax in South Africa. It declined to make firm recommendations in view of the complexity of its administration and the lack of capacity of the Inland Revenue at that time.

This submission seeks to place the capital gains tax in context. Many of the issues raised are the subject of ongoing policy analysis and difficult policy trade-offs as the capital gains tax legislation is being prepared.

2. Capital Gains Tax and the Fiscal Framework
Since 1994 the South African fiscal authorities embarked on systematic structural adjustments, thereby rectifying a series of macroeconomic imbalances. These fiscal reforms translated into a marked reduction of inflationary pressures and fiscal deficits. In line with the macroeconomic adjustments, real GDP growth improved to an average of 2.3 per cent per year over 1995 to 1999 as compared to 0.2 per cent over 1990 to 1994.

In future years, the main macroeconomic and fiscal challenge remains to increase economic growth in order to address the serious unemployment, poverty and related crime problem. Needless to say, robust growth needs to be underpinned by increases in savings and investments.

Fiscal restraint translated into National Government’s systematic reduction of the deficit before borrowing from 5 per cent of GDP in 1994/95 to approximately 2 per cent for 1999/2000, and rising slightly to an estimated 2,5 per cent for 2001/02. This impressive improvement can largely be ascribed to strong revenue collection performance, which was backstopped by attractive and effective tax reform initiatives and improved tax administration. Indeed, the combination of these policy prompts was singularly aimed at improving the efficacy and efficiency of the South African tax system.

The tax reform agenda was initiated by the new Government and was executed under the able leadership and wisdom of the Katz Tax Commission. The Commission’s work was largely informed by the global paradigm shift of eliminating tax incentives or preferences throughout the tax codes, thereby enabling tax base broadening that could finance a marked reduction of tax rates in the field of the personal and corporate income tax systems. The principle behind the attractiveness of a low rate tax system is that it makes tax-planning activities less lucrative, thereby freeing up scarce human resources in any economy. Moreover, an overall lower rate structure would address effectively distortions in the economy that result from certain sectors enjoying huge tax privileges, which per definition must lead to an over-investment in those sectors as investment risks are artificially being lowered.

2.1 Composition of national taxes
The standard corporate tax rate was reduced from 40 to 30 per cent (15 per cent for small and medium-sized enterprises, whilst personal income tax benefited from rate reductions, scrapping of tax brackets and by reducing the compression within the brackets.

Table 1 Composition of national tax revenue

The Fiscus realised its revenue targets with ease. National Government or Main Government Revenue is estimated to amount in 2000/01 to R216,4 billion or 24,1 per cent of GDP. Taxes on income and profit are responsible for almost 60 per cent of the total revenue take of which individual income tax contributes close to 41 per cent of total revenue, the single most important revenue source for Government. It is important to note that the revenue importance of corporate income tax in total tax revenue evidences a long-term decline to approximately 10 per cent in 2000/01, whereas it used to generate up to 17 per cent of total tax revenues in 1983/84.

The introduction of capital gains tax therefore will seek to adjust a fundamental weakness in the South African tax system as it will minimise the significant arbitrage opportunities that exist for avoiding tax through the recharacterisation of ordinary taxable income or revenue streams into untaxed capital gains. As corporations, especially the sophisticated players in the local financial markets, are commonly the well-advised taxpayers it is obvious that the introduction of capital gains tax will seek to reduce the high tax burden on individuals and to shift the tax burden more equitably across the different categories of taxpayers. As pointed out in the IMF Report, this is an additional, necessary and long overdue base-broadening exercise, which eliminates the current zero-rated income tax on capital gains income.

3. Rationale for taxing capital gains
Economic policy often involves important trade-offs. Tax policy is no exception. As Stiglitz (1988, pp 478-479) notes:
"The optimal tax structure is one that maximises social welfare, in which the choice between equity and efficiency best reflects society’s attitude toward these competing goals."
The introduction of a capital gains tax is most often justified on fiscal equity grounds, but other criteria are also important. International best practice strongly suggests that capital gains should ideally attract the full income tax as other forms of income as the continued preferential treatment of capital gains will perpetuate avoidance and decline in tax morality by the low income taxpayer who cannot afford the use of tax advisers. It widens the income tax base, secures the existing base by limiting tax avoidance activities, improves equity and reduces investment distortions.

3.1 International trends and practice in applying the income tax on the basis of the comprehensive income concept
Many jurisdictions accept the soundness of the ‘comprehensive income’ concept as the ideal tax base. This approach was outlined by Haig and Simons and means that the total sum of all revenue streams over the tax period should be included in the income tax base as it constitutes increases in the purchasing power of a taxpayer. Capital gains represent one of these income streams and should attract the same income tax charge as all other revenue streams. Taxing capital gains is therefore relatively common internationally. In 1996, New Zealand was the only OECD country that did not tax capital gains explicitly, though as Oliver (2000: 2) notes:
"[I]t seems a bit simplistic to describe a tax system as one that does or does not tax capital gains. Any income tax that left all capital gains tax-free would be unworkable".
In reaction to the circulated Draft Taxation Law Amendment Bill on the introduction of capital gains tax provisions, the argument was put forward that the introduction of capital gains tax provisions would be impractical and inappropriate in the case of developing countries and went even further by suggesting that indeed many developed countries are in the process of phasing out capital gains tax. However, numerous cross-country comparative studies clearly refute these claims. Table 2 outlines the respective CGT regimes in both developed, middle income and developing countries. The analysis is based on the most recently published and comparable tax legislation information available.

Table 2: Cross-country review of CGT systems - AFRICA

Table 2 continued: ASIA

Table 2 continued: AMERICA

Table 2 continued: CARIBBEAN AND OTHER COUNTRIES IN ASIA (MIDDLE EAST)

Sources:
PriceWaterhouseCoopers (1999) "Corporate Taxes: Worldwide Summaries1999-2000", John Wiley and Sons, Inc (USA)
PriceWaterhouseCoopers (1999) "Individual Taxes: Worldwide Summaries1999-2000", John Wiley and Sons, Inc (USA)
Kesti, J and Balle, C.H. eds (2000) European Tax Handbook, IBFD, Amsterdam
International Bureau for Fiscal Documentation, 1998-2000. African Tax Systems, IBFD, Amsterdam.
International Bureau for Fiscal Documentation, 1998-2000. Taxation in Latin America, IBFD, Amsterdam.
International Bureau for Fiscal Documentation, 1998-2000. Taxes and Investment in Asia and the Pacific, IBFD, Amsterdam.
International Bureau for Fiscal Documentation, 1998-2000. Taxation of Individuals in Europe, IBFD, Amsterdam.

In sum, and based on limited information available it would appear that –
i. In the case of Africa, 14 out of 43 jurisdictions have decided against the introduction of capital gains tax or data is simply not available, which is 32 per cent, whilst 68 per cent of the jurisdictions opted for some form of inclusion of capital gains.
ii In the case of Asia and the Asian Pacific region, 21 jurisdictions or 46 per cent decided against capital gains tax, 54 per cent elected to tax capital gains but sometimes on only a very limited asset class, such as real property. In some cases marketable securities or shares of listed companies are exempt at the level of natural persons.
iii. In the Americas, 2 out of 19 jurisdictions (Costa Rica and Ecuador) have excluded capital gains tax provisions from their respective income tax legislation (11 per cent) – the others all tax capital gains (89 per cent). However, it is interesting to note that Ecuador introduced a financial transaction tax of 1 per cent on turnover as a presumptive tax measure for the financial service sector. Argentina exempts gains realised by individuals on shares, bonds and other securities.

ii. 12 jurisdictions (55 per cent) in the Caribbean and the Middle East opted against the introduction of capital gains tax provisions. It needs to be noted that many of these are oil-producing countries saw hitherto no need to introduce an income tax system.
iii. In Europe 31 jurisdictions saw the need to include realised capital gains into the income tax system.

4. Capital gains tax and equity
Tax policy and tax reform are informed by the three central tenets of tax design – equity, efficiency (including inter-sectoral neutrality) and simplicity. It is against these standards that the capital gains tax should be evaluated, which is the focus of the remainder of this section.

According to the comprehensive income definition, capital gains should be treated no differently from other forms of income. Shome (1995: 7) defines comprehensive income as:
"… the sum of the market value of rights exercised in consumption and the change in the value of the store of property rights between the beginning and the end of the period in question. Thus, this definition of ‘comprehensive’ income equals consumption plus net wealth accumulated during the period".
Equity in taxation consists of both horizontal and vertical equity. Horizontal equity demands that individuals in similar economic circumstances should bear a similar tax burden, irrespective of the form the accretion of economic power takes. In other words, taxpayers should bear similar tax burdens, irrespective of whether their income is received in the form of wages, or capital gain. In this context, the exclusion of capital gains from the income tax base fundamentally undermines the horizontal equity of the tax system.

Vertical equity connotes that taxpayers with greater ability to pay taxes should bear a greater burden of taxation. It is common cause that capital gains accrue disproportionately to higher income individuals. Thus, including capital gains in taxable income contributes to the progressivity of the income tax system, while enabling government to pursue other tax policy objectives, premised on widening tax bases and reducing standard tax rates.

Furthermore, international experience indicates that the biggest share of capital gains tax revenues can be attributed to the wealthiest of individuals. As was indicated earlier, in Canada and the United States the top 1 per cent of taxpayers generate approximately 60 per cent of the capital gains. The high inequality in South Africa, with an income Gini coefficient of 0.61 as reported by Aron and Muellbauer (2000) makes it absolutely necessary to include capital gains.

For illustrative purposes Table 3 (Burman 1999:114) details the American experience that the distribution of capital gains tax liability is more heavily concentrated at the high income levels than the distribution of realised capital gains across the taxpayer population. Table 3 indicates that taxpayers who had incomes of more than $200 000 realised 61.9 per cent of all capital gains in the United States and paid 62.4 per cent of the tax. Given the present exclusion threshold of R10 000 per annum for the proposed South African capital gains tax system, it appears to be logical that a similar pattern of tax liability concentration within the ranks of high-income earners will be found. The introduction of capital gains tax will therefore markedly improve the vertical equity of the income tax system in South Africa.

Table 3: Distribution of taxes paid on capital gains in the United States, by income, 1993
In Millions unless otherwise specified

4.1 Capital gains and economic efficiency
Excluding capital gains from the income tax base is an important structural weakness in the income tax system, which leads to tax avoidance and the misallocation of productive investment resources. The absence of a capital gains tax encourages taxpayers to convert ordinary taxable income into tax-free capital gains, thereby reducing their total tax burden. The application of scarce resources to tax planning and tax avoidance is clearly a dead-weight loss to society.

The efficiency case for introducing a capital gains tax is particularly strong if one considers the impact on the allocation of investment funds. If capital gains go untaxed, individuals are encouraged by the tax system to invest their savings in assets that provide returns in the form of capital gains (e.g. property), rather than income producing assets (e.g. equipment and machinery). Scarce investment funds are clearly misallocated when tax factors are given undue weight over risk-return considerations in the allocation of investment capital. Capital gains tax will narrow the gap in the tax treatment of different assets, reducing these distortions in individual portfolio decisions.

Further efficiency arguments are explored in Section 5, where the impact of capital gains tax on economic activity is considered.

4.2 Capital gains tax and simplicity
Simplicity of taxation requires that the tax system should not be overly difficult for the revenue authorities to administer and should not impose onerous compliance burdens on taxpayers. A capital gains tax is difficult to justify on the grounds of simplicity – it is, in whatever form it is implemented around the world, a rather complex tax. However, as Oliver (2000: 3) notes, not taxing capital gains does not necessarily lead to a simple tax system:
"[s]implicity has not been the outcome of a lack of capital gains tax in New Zealand".
As in South Africa, the income tax in New Zealand has become very complicated because a number of anti-avoidance provisions and other legislative solutions have been necessary to address tax avoidance schemes.

Furthermore, careful design of the capital gains tax can avoid some of the complexity inherent in the tax, thereby reducing the administrative and compliance burdens.

4.3 The capital versus revenue distinction - could more certainty be created?
The Katz Commission Report properly noted that the capital gains debate involves two related issues: (1) whether capital gains should be taxed, and (2) whether capital gains should be subject to the same rate of tax as ordinary revenue. As previously stated, the National Treasury is committed to the taxation of capital gains as part of its base broadening efforts and on grounds of equity. After careful consideration, the National Treasury has reluctantly decided not to tax capital gains at ordinary rates despite the existence of strong arguments to the contrary. Some of these contrary arguments include:
i. Horizontal Equity, Vertical Equity, and Taxpayer Morale: Profits from the sale of capital assets have the same economic value as profits from wages and other ordinary revenue. In terms of horizontal equity, fairness demands that capital gains should be subject to the same rate of tax. In terms of vertical equity, capital gains are mostly generated by the wealthiest taxpayers. Failure to tax these gains at ordinary rates accordingly undermines the goal of progressivity. Violations of these forms of equity ultimately leave the ordinary wage earner subject to a higher level of tax, thereby undermining their notion of fairness and taxpayer morale.

ii. Complexity and Avoidance: As long as a differential exists between capital gain and ordinary rates, taxpayers will have an incentive to characterise their gains as capital. A complex set of case law has emerged attempting to define this distinction. Taxpayers also have a long history of engaging in transactions that convert capital gains into revenue (for instance, taxpayers once disguised bond interest though bonds that yielded no stated interest, disguising this interest through higher redemption values). According to the Commissioner, one in five cases litigated by SARS involves the capital versus ordinary distinction.

iii. Economic Efficiency: As long as a differential exists between capital gains and ordinary rates, taxpayers have an artificial incentive to convert capital gains into ordinary revenue, thereby undermining economic efficiency. The result is a tax system that effectively encourages otherwise uneconomic or sub-optimally efficient transactions.

iv. Sharing Downside Risk and Counter-Cyclical Stimuli During Economic Downturns: Under a preferential system of capital gains, capital losses only offset capital gains (i.e., income taxable at a 10.5/15 per cent). In a world of full inclusion, capital losses would become valuable because these losses would offset ordinary revenue (i.e., income taxable at a 42/30 per cent). Full allowance of loss would mean the government essentially becomes a larger partner in taxpayer loss, thereby further encouraging investment in risky ventures. Full allowance of loss would also provide taxpayers with an automatic indirect subsidy during periods of economic downturn. The result would mean an automatic counter-cyclical stimuli for recovery.

These arguments against preferential treatment are essentially the same as the arguments against the current wholesale exclusion. However, consistent with international practice, the capital gains legislation proposed utilises a lower rate for capital gains (the maximum 10.5 per cent for individuals and the flat 15 per cent rate for corporations). The National Treasury ultimately decided that the change from a zero rate to a full inclusion rate was simply too disruptive in the current business environment and could dissuade foreign investment in an international environment where capital gains are typically subject to preferential treatment.

Admittedly it is most unfortunate, that the incentive for tax arbitrage between capital versus revenue will continue, but this arbitrage will be reduced with capital gains become partially taxable. This partial taxation reduces the current differential between capital gains versus ordinary revenue. More importantly, the proposed system of partial taxation will mean that taxpayers must report their capital gains for the first time, thereby eliminating evasion. Taxpayers can no longer simply hide their gains from SARS by describing their gains as capital. The enhanced reporting requirements will simultaneously enhance the National Treasury's understanding of taxpayer behaviour in reaction to tax amendments, which should improve policy makers' ability to model more accurately behavioural shifts.

The bunching problem
In making its decision to utilise a preferential rate, the National Treasury also took into account concerns about inflation and bunching of income within a single year. As previously stated, a gain on the sale of capital assets includes a significant inflationary element. The sale of capital assets often has the further effect of triggering high levels of gain within a single year, even though the gain on the asset may have accumulated over multiple years. These points bear some force. However, these arguments are not without counter-argument. All forms of income contain some inflationary element, and the Income Tax Act typically contains no concession. Capital gain assets additionally have the benefit of deferral. Unlike wages, which are taxed annually, the tax on capital gain does not arise as the gains annually accrue; the tax on these gains arise only upon disposal. International studies illustrate this benefit of deferral ultimately outweighs the detrimental inflationary and bunching effect.
Safe harbour rules for distinguishing between ordinary revenue versus capital gains
The proposed capital gains regime leaves the current gains versus ordinary revenue distinction of case law fully in place. Written submissions to SARS rightly indicate that this case law distinction is vague, creating taxpayer uncertainty as well as needless litigation.

The National Treasury is also concerned that the present distinction between capital gains versus ordinary revenue makes little economic sense. In summary form, case law treats the sale of an asset as ordinary revenue if a taxpayer holds an asset with the intent to sell. Case law conversely treats the sale of an asset as capital gain if a taxpayer holds the same asset with the intent to produce income. This intent test is based on the taxpayer’s stated intent along with objective external factors evidencing that intent (such as the taxpayer’s history with respect to similar assets).

Take the following example: If a taxpayer purchases shares with the intent to sell those shares at a gain, the gain generates ordinary revenue. If a taxpayer purchases shares mainly for the dividend yield, the gain is capital.

As the above example illustrates, the distinction between capital gains and ordinary revenue has no economic rationale. Why should the taxpayer’s intent matter? Moreover, can one really argue that a taxpayer purchases shares for dividends (other than certain preferred shares) in today’s market of share speculation. The National Treasury further understands that financial institutions typically are of the view that their sale of shares can be capital gain based on the same distinction. We sincerely doubt whether sophisticated institutions such as these can seriously claim they hold shares merely for dividends in today’s sophisticated economic climate.

In sum, the National Treasury is currently undergoing a study of this distinction along with recommendations for legislative reform. The proposed distinction will provide for a stricter set of objective guidelines rather than the current test based on intent. The Treasury is focusing on a number of criteria such as whether the gain on similar assets has become akin to salary for the taxpayer involved and the time period for which the asset at issue is held.

5.Structure of proposed capital gains tax
In an ideal comprehensive income tax system, only real, non-inflationary capital gains that accrue to a taxpayer over a tax period would be taxed. It needs to be emphasised that from a theoretical perspective capital gains tax should be taxed as fully as other forms of income or profits, and then most definitely from capital.

Any tax preferential treatment of capital gains -- which is the case in the currently proposed capital gains tax provision -- will still create the same opportunities for tax arbitrage and avoidance which will be fully exploited by the well-advised and wealth taxpayers in South Africa. From an economic perspective, it will still be inefficient to employ highly skilled and scarce legal and accounting resources to exploit the arbitrage opportunities, which will add to the pressure of constantly designing new anti-avoidance measures, thereby adding to the complexity of the current draft legislation. This was eloquently summed up in the IMF Report, which lamented the fact that concessionary treatment of capital gains does not free up resources in the tax planning and administration community, that transaction cost are not adequately reduced as complex anti-avoidance measures need to be designed in order to arrest new avoidance schemes.

In this regard one would like to state that the National Treasury’s preference lies in the minimal use of tax preferential treatment of capital gains, that ideally all revenue accruals should be taxed, that there should be a full inclusion rate for individuals, special trusts and companies. Given the current fiscal policy stance of maintaining a tax-GDP ratio of 24 to 25 per cent, this significant base broadening exercise would enable the Fiscus to reduce the overall personal and corporate income tax rate structure. As rates would come down, the need for tax planning would be minimised. Differently put, as demands would mount for lower inclusion rates, further exemptions for personal use assets, rollover relief for business reinvestments, preferential treatment to gains realised at death or when gifts are made the rate structure for the personal and corporate income tax system needs to be adjusted upwards, which creates further unintended distortions and dead-weight loss for the economy.

The realities of tax policy are however such that no country has implemented capital gains tax in this pure form. The "Guide to Capital Gains Tax" released at the time of the budget proposed the following key design elements for the capital gains tax.

i. Capital gains tax will be levied on South African residents in respect of their worldwide assets and on non-residents on gains arising out of the disposition of immovable property and gains accruing to a permanent establishment, which is in line with international norms.
v. Certain assets, including primary residences, will be exempt.

6.Capital gains tax and economic activity
Critics of capital gains tax often suggest that taxing capital gains will undermine economic efficiency and have deleterious effects on South Africa’s economic performance. These include charges that a capital gains tax will reduce the level of savings, investment and economic growth and impinge on risk-taking and entrepreneurial activity. General comments regarding the efficiency effects of taxing capital gains are set out in Section 4.2. This section focuses specifically on the impact on saving, investment and economic growth.

6.1 Capital gains tax and saving
ii. Capital gains tax will be levied on all gains realised after the "effective date" (1 April 2001), but only on the portion of the gain accruing after the effective date. Taxpayers can elect to have assets valued on 1 April 2001 or could rely on ‘time-based apportionment’ to ensure gains accruing before the effective date are not taxed.

iii. Individuals will include 25 per cent of net capital gains and companies will include 50 per cent of net capital gains in taxable income, which will be taxed at ordinary income tax rates. As a result, the effective capital gains tax rates for individuals will vary from 0 - 10,5 per cent and the rate for companies will be 15 per cent.

iv. The capital gains will not be indexed for inflation. The combined benefits of the ‘low inclusion rate’ and deferring accrued capital gains until realisation should more than compensate for the effects of inflation in a moderate-inflation environment. The inflation rate has fallen from an average of 12,5 per cent between 1990 and 1994 to 7,4 per cent between 1995 and 1999. The inflation-targeting regime currently in place will further contribute to reducing inflation.

National saving in South Africa has fallen from 27,1 per cent of GDP in the period from 1979 - 1984 to 16,1 per cent of GDP for 1994 – 1998. For sustained investment and economic growth, it is imperative that this falling trend is reversed. It might be asked whether introducing a capital gains tax will contribute to reversing this trend or aggravate it. The answer to this is not straightforward and involves considering the impact on both household saving and government saving.

Theoretically, the effect of taxation on household saving is ambiguous, due to offsetting income and substitution effects. The empirical evidence on the effects of taxation on the level of household saving is also not compelling. As suggested in an OECD study:
"…there is no clear evidence that the level of taxation … does generally affect the level of household saving." OECD (1994).
Burman (2000: 57) further notes:
"…changes in the taxation of capital gains are likely to have little effect on private saving".
The absence of a capital gains tax in the past does not seem to have influenced the level of saving much, if one considers that South Africa’s savings rate is lower than many countries that do have taxes on capital gains. It is likely, therefore, that other factors (e.g. level of lifetime income, demographics) have a greater effect on the level of household saving than do taxes.
In this context, one must also consider the positive effect of capital gains tax receipts on government revenues. While the direct receipts tend to be small, the total contribution of a capital gains tax is greater than the direct revenue flow. By reducing the incentives to convert ordinary taxable income into tax-free capital gains, the capital gains tax enhances the ordinary income tax base, increasing receipts from this source as well. These receipts can then be applied to reducing government dissaving. Alternatively, the resources can be used to reduce statutory income tax rates, which will contribute to the efficiency of the overall tax regime. In sum, the introduction of capital gains tax could well have a positive effect on the level of national savings.

6.2 Capital gains tax, investment and economic growth
Sustained economic growth requires considerably more investment than the current levels. Will the introduction of a capital gains tax undermine other efforts to increase the level of investment? The impact of capital gains tax on investment and economic growth must be traced through three steps:

i. The effect of the capital gains tax on the cost of capital (the rate of return an investment project must offer to attract investment funds).
ii. The elasticity of investment spending with respect to changes in the cost of capital. This measures the behavioural response of firms and investors to changes in the cost of capital.
iii.The effect of investment on economic growth.

The user cost of capital has two components – the economic depreciation rate of the tangible capital asset and the financial (opportunity) cost of the investment, which measures the financial returns foregone by sinking money into fixed capital, rather than financial assets, adjusted for inflation.

Capital gains tax only affects the financial cost of capital. It is, however, not the sole factor that affects this variable – it is also influenced by real interest rates, inflation rates, tax rates on interest and dividends, corporate tax rules (including tax depreciation rates, investment allowances and basic corporate tax rates). It is beyond the scope of this paper to analyse these factors in detail, but there is little reason to think that the introduction of capital gains tax will significantly affect the cost of capital. This is particularly true for assets where the economic depreciation is a significant proportion of the total cost of capital, as in the case of assets with short economic lives.

One must also consider that domestic saving is not the only source of investment funds. In the South African context, domestic savings are supplemented by capital inflows from the rest of the world, which are largely unaffected by South African capital gains taxes.

There is substantial uncertainty in the economics literature regarding the responsiveness of investment spending to changes in the cost of capital, yet it is a critical parameter in establishing the effects of taxes on investment. In the popular debate on the effect of taxes on investment, it is often assumed that there is a one-for-one response, i.e. a 1 per cent increase in the cost of capital would reduce investment spending by 1 per cent. Fazarri (1999: 3) questions this assumption, submitting that it could be much lower, quoting the estimate of 0,25 (a 1 per cent increase in the cost of capital would reduce investment spending by 0,25 per cent) derived by Chirinko, Fazarri Meyer (1999).

The third channel considers how increased investment affects economic growth. Most studies assume that a 1 per cent increase in the capital stock would increase output by about 0,3 per cent.

In sum, the effect of capital gains tax is a problem of multiple elasticities – the change in the cost of capital resulting from the capital gains tax; the change in investment resulting from the change in the cost of capital; and the change in economic output resulting from the change in investment. As Fazzari (1999: 4 notes):
"The problem of multiplied elasticities comes down to the fact that when three numbers much less than one are multiplied together, the resulting number is quite small."
In this context, there is no reason to assume a large impact of capital gains tax on economic growth.

6.3 Risk-taking
Capital gains tax is often argued to limit risk taking and entrepreneurship. A tax on capital gains lowers the after-tax return from a high-risk project. However, by allowing deductions for capital losses, the potential loss from the project is also reduced, which reduces the variability of returns from any high-risk investment project, making these investments more attractive. While this advantage is reduced because capital losses are not set off against other income, most investors in high-risk projects have diversified portfolios and are able to offset capital losses. In sum, the ultimate effect of capital gains tax on risk-taking depends on a number of factors, including inter alia:

i. the risk profile of the investment;
ii. the attitude of the investor toward risk (i.e. the degree of risk aversion); and
iii. the extent to which the investor’s portfolio is diversified (so any capital losses could be offset against other taxable gains)
.
7. Capital Gains Tax: INFLATION INDEXING
Capital gain income falls within the Schanz-Haig-Simons definition of income and, as such, should be treated no differently from other income in a comprehensive income tax system. Shome (1995: 7) defines this "comprehensive income" as:
"… the sum of the market value of rights exercised in consumption and the change in the value of the store of property rights between the beginning and the end of the period in question. Thus, this definition of ‘comprehensive’ income equals consumption plus net wealth accumulated during the period".

An appreciation in the real value of assets (for example, real property or financial assets such as company shares) implies a change in the net worth (wealth) of the individual holding those assets. In line with the comprehensive income definition, this increase in wealth should be included in taxable income. Allowing capital gains to remain untaxed compromises the principle of horizontal equity, as two taxpayers in similar economic circumstances would bear different tax burdens. Furthermore, since capital gains generally accrue disproportionately to wealthier members of society, introducing a capital gains tax will improve the vertical equity of the overall tax system, as those who realise capital gains will (in line with the "ability to pay principle") pay a fairer share of income tax. Given South Africa’s high income inequality, introducing capital gains tax is absolutely vital to ensuring an equitable distribution of the tax burden.

The efficiency of the tax system is improved as incentives for taxpayers to convert taxable ordinary income into untaxed capital gains are reduced by the imposition of a capital gains tax. Given the sophistication of South Africa’s financial markets, it is relatively simple for well-advised taxpayers to exploit the capital / revenue distinction inherent in the income tax law. While the proposed capital gains tax will not eliminate these arbitrage opportunities, it will certainly reduce them.

The imposition of a capital gains tax results in more revenue for the Treasury in two ways. The capital gains tax may raise revenue in itself or, by reducing the opportunity for converting ordinary taxable income into untaxed capital gains, there may be increased collections elsewhere under the income tax.

As with all tax policy and tax reform efforts, the introduction of a new tax involves balancing the central tenets of a good tax system – efficiency, equity and simplicity. The theoretically correct approach to taxing capital gains, in the spirit of the Schanz-Haig-Simons definition of income, would be to include all real capital gains in taxable income together with other ordinary income, as they accrue to the taxpayer. However, while most countries recognise the importance of taxing capital gain income, none have implemented a theoretically pure capital gains tax. In practice, important compromises are made in the interests of administrative simplicity.

First, although economic theory points to accrual taxation most countries operate a capital gains tax on a realisation basis. Usually motivated by concerns that taxpayers would not have sufficient liquidity to meet the annual capital gains tax obligation, capital gains taxes are levied on a realisation basis. In other words, the capital gains tax is deferred until taxpayers realise such gains. Also, it is hard to assess value for illiquid assets such as real estate or businesses other than when they are traded. Italy is an exception to this. The Italian tax system taxes most capital gains on mutual funds and other assets managed by financial institutions on an accrual basis.

The second area in which practical compromise results in a deviation from the theoretically pure application of capital gains tax is in the area of inflation indexation. As has been noted, only real capital gains are an "accretion" to a taxpayer’s wealth and should be subject to capital gains tax. It is argued that taxing nominal capital gains can result in "illusory" gains being taxed because unindexed capital gains tax can lead to effective rates of tax on real capital gains that deviate from the statutory rate.

The central question however, is whether ad hoc adjustment mechanisms that mitigate the impact of inflation in capital gains tax regimes and considerably reduce the administrative complexity of such systems, are reasonable compromises. If other income (e.g., rent, interest and dividends) is not indexed, it is no longer obvious that capital gains should be indexed. Taxing capital gains on a real basis while other income and expenses are taxed on a nominal basis creates inefficiency, as well as inequity.

The discussion document released on 23 February 2000 by the National Treasury and the South African Revenue Service outlining the framework for the proposed capital gains tax regime, indicated that the system would not include inflation indexation. This aspect of the discussion document attracted significant attention.

This is the main focus of this submission, which considers:
i. The general effects of inflation on tax systems.
ii. The advantages and disadvantages of inflation indexing.
iii. The impact of inflation in the proposed South African capital gains tax regime.
iv. Country experiences with inflation and the extent to which capital gains tax systems are indexed for inflation.
v. The appropriate policy response in the current South African context.

From the discussion set out in this document, it is proposed that South Africa’s current inflation position and the future inflation outlook do not necessitate the administrative complexity of introducing systematic inflation indexing in the capital gains tax system. The regime, as it has been set out with reduced capital gains tax inclusion rates, on a realisation basis, will more than adequately compensate for inflation in the system. In fact, at realistic assumed rates of inflation and moderate real rates of return, the effective capital gains tax rates under the proposed system are lower than under a fully indexed system.

7.1 IMPACT OF INFLATION ON TAX SYSTEM
7.1.1. General
Inflation can affect virtually every aspect of the tax system, including the personal and corporate income tax. Thuronyi (1996) suggests that inflation can affect the real tax liability in the following ways:
i. Eroding amounts expressed in national currency, e.g. depreciation allowances based on historical cost.
ii. Eroding the value of tax obligations if there is a delay between when the "tax event" occurs and when the payment is made.
iii. Causing mis-measurement of the tax base, e.g. valuation of inventory, etc.

7.1.2 Impact of inflation on the Personal income tax
On the personal income tax, inflation has two effects, namely, increasing the taxable income of individuals in nominal terms (thereby increasing the effective tax rate on real income) and eroding the value of deductions/exemptions.

In a personal income tax system based on graduated marginal tax rates, inflation can cause "bracket creep". A nominal increase in income, which maintains a constant real income could result in individuals moving into a higher tax bracket and paying a higher marginal (and average) tax rate on their income, even though their real income is unchanged. In periods of high inflation, this "inflation tax" can become quite severe. Secondly, inflation erodes the real value of deductions/exemptions that remain fixed in nominal terms over time, e.g. monetary caps on the contributions to retirement annuities.

7.1.3 Impact of inflation on the Corporate income tax
In the corporate income tax system, there are many areas that are affected by inflation, including:
i. The value of interest deductions are increased, as part of the interest is actually compensation for inflation, with the result that companies are able to deduct part of the capital repayment in the income tax system.
ii. The value of interest income that is subject to tax is increased by the inflation component of the total interest charge.
iii. The value of depreciation allowances, which are based on historical cost, are eroded over time with inflation. The impact of this obviously depends on the length of the depreciation period and the inflation rate.
iv.In systems where inventories are valued according to the FIFO (First-In-First-Out) method, there could be an "inflation tax" when inventories are revalued.

In fact, Thuronyi (1996: 442) states the following:
"On a more abstract level, inflation can be seen as destroying the integrity of all forms of accounting, including tax accounting, based on historic costs. Inflation makes it impossible to add or subtract amounts such as receipts, expenses, inventory balances, and so on, stated at historic costs and occurring on different dates. It is as if these numbers were expressed in different currencies. At high levels of inflation, the tax base becomes virtually meaningless." (emphasis added).

It is certainly true that high rates of inflation can seriously undermine the integrity of the tax system. That is, the entire tax system not just parts of it, e.g. the capital gains tax. However, one must consider what is meant by high rates of inflation and when indexing of all or part of the tax system is warranted. The pro’s and con’s of inflation indexing are discussed more fully in section 7.2.

7.1.4 Impact of inflation on the Capital gains tax
Because capital gains are taxed on a realisation basis, as opposed to an accrual basis, it is often submitted that inflation can have a much greater impact on the capital gains tax system than the ordinary income tax system. In essence, the capital gain that is subject to tax on disposal of an asset, is the difference between the acquisition cost and the selling price of the asset. It is argued that because there could be a considerable time period between the acquisition and disposal of an asset, i.e. before a capital gain crystallises, the impact of inflation is more acute in the capital gains tax regime. This is not true. The effective tax rate on capital gains declines with the holding period of the asset, conferring a tax deferral benefit to the owner of the asset. Effectively, rent, interest and dividends are taxed at full rates every year. Even if capital gains were fully taxed upon realisation, their effective tax rate would be lower than the rate on other capital income on assets held for more than one year.

7.1.5 Methods of adjusting for inflation
Countries that have found it necessary to introduce specific measures into their tax system to take account of inflation, have used basically three methods:
i. Global adjustment, where the effects of inflation are comprehensively removed from the tax system. According to Thuronyi (1996: 465):
"The global adjustment method is used in practice only by countries with relatively high inflation – 30 percent or above – or by countries that have had such levels in the past and have consequently already put global adjustment into place."

ii. Global adjustment has been practised for some time in countries that have suffered from chronically high inflation, including Argentina, Brazil, Chile, Colombia, Mexico, Peru, Venezuela, and Israel. Romania adopted this method for its profit tax in 1994.
iii. Partial adjustment involves adjusting for inflation with respect to particular items of income or deduction, usually by indexing the base cost of capital items. Interest income or expense may also be adjusted for inflation.
iv. Ad hoc adjustment involves applying measures that are not explicitly based on calculating the amount of inflation, but are designed to offset the rate of tax on capital income. It is submitted that at low and stable rates of inflation, ad hoc adjustment methods can function fairly well. Examples include:
-Applying a lower rate of tax on capital income
-A partial exclusion of capital gains income

There are generally two kinds of partial exclusion systems, a fixed percentage exclusion regardless of the length of the asset holding period and variable percentage exclusion which increases with the asset holding period.

The United Kingdom has replaced inflation indexing for individuals with variable percentage exclusion, called taper relief, for disposals of assets on or after 6 April 1999. The United Kingdom’s taper relief reduces the amount of the chargeable gain according to the length of time the asset has been held after 5 April 1998. For example, for non-business assets, the gain is reduced by 5 per cent for each complete year of ownership after the first two years of ownership; the maximum reduction is reached after at least 10 years of ownership when only 60 per cent of the gain is taxable. The business asset taper now runs as follows:

It should also be borne in mind that capital gains tax systems that only tax realised capital gains, instead of accrued capital gains, also provide a benefit to taxpayers that can also in some way mitigate the effects of inflation.

"The realisation rules of the income tax can also act as an ad hoc offset to inflation. At certain holding periods and inflation rates, the benefit to the taxpayer of being able to defer taxation of the gain until realisation roughly offsets the detriment of having to pay tax on the inflationary component of the gain." (Thuronyi, 1996: 444).

7.2 Pro’s and con’s of INDEXING CAPITAL GAINS tax
7.2.1 Advantages of indexation
Consistent effective tax rates
If capital gains taxes are fully indexed for inflation, there will be consistency between the effective capital gains tax rate and the statutory tax rate. Indexing reduces the uncertainty arising from inflation. If capital gains are not indexed for inflation and inflation turns out to be surprisingly high, taxpayers pay higher real effective tax rates.

In systems where nominal capital gains are taxed, the effective tax rate will diverge from the statutory rate and the degree of such divergence will depend on the inflation experience (as well as the interaction between the inflation rate and the real rate of return). This will be explored further in section 7.3, where it is outlined that for a constant inflation rate and real rate of return, the effective tax rate falls over time (i.e. as the period for which the asset is held increases). However, where the inflation rate is not constant over time, the effective tax rate will depend on the actual inflation experience, and will be somewhat arbitrary. Needless to say, this compromises one of the central objectives of tax policies – certainty.
"Inflation in an unindexed tax system also tends to create a haphazard pattern of effective tax rates in the economy that may distort the market’s ability to efficiently allocate resources" (Thirsk, 1997: 14).
Distortion of investment incentives
It is argued that the absence of indexation can distort investment decisions when inflationary gains can potentially be subject to taxation. It is suggested that not indexing capital gains tax would discourage investment in long-term assets as the uncertainty regarding the effective capital gains tax rate on disposal of the asset is increased with the holding period.

Both indexation and partial exclusion of nominal gains create a tax arbitrage on various types of assets. Both favour growth assets over dividend-paying assets because retained earnings enjoy a tax deferral benefit that increases with the length of the asset holding period. For example, dividends are taxable in the year they are declared (Secondary Tax on Companies) and thus cannot enjoy the tax deferral benefit. Given these distortions, it has been suggested that indexation is a more efficient way to grant such a tax preference to growth assets than a partial exclusion.

"Among depreciable assets that pay capital gains, indexing of structures would be more favourable than an exclusion." (Congress of the United States: Congressional Budget Office, August 1990: 13).

7.2.2 Drawbacks of indexing capital gains tax
Investment distortion
It has been noted that investment decisions can be distorted if the capital gains tax system is not indexed for inflation. However, this must be seen in the wider context of the entire tax system. Only indexing one part of the tax system will result in further distortions in the level and pattern of investment, as well is in the financing of that investment.
"When an indexed treatment, however, is introduced for capital gains but not for other forms of income, such as interest income, whose measurement is similarly affected by inflation, new anomalies and distortions can arise in the tax system." (King in Shome (1995: 158).
It was noted that interest on debt consists of two components – one that compensates for the costs of borrowing and another that compensates for inflation. When nominal interest costs are deductible, investors are encouraged to finance investment through debt as opposed to other financing mechanisms (i.e. equity). If the capital gains tax is fully indexed and other parts of the income tax system are still based on nominal values, investors would be encouraged (artificially by the tax system) to invest in capital growth assets and finance such investment through debt. While these distortions are not eliminated under the proposed capital gains tax (because of the low inclusion rates and deferral benefits), they is certainly reduced.

Loss-limitation rules can also impact on the efficiency and effectiveness of indexation of capital gains tax:

"…depend on how indexing would be done. Partly because of concern about revenue losses, most indexing proposals … call for a special limitation on losses created by indexing. The limitation would allow nominal losses (unadjusted for inflation) to be deducted from net capital gains … but would not allow the deduction of losses caused solely by inflation. As a result of the limitation, indexing could never convert a nominal gain into tax loss, regardless of how much inflation had eroded the purchasing power of the original investment." (Congress of the United States: Congressional Budget Office, August 1990: xiii).

Further, as Burman notes:
"For assets with nominal gains but real losses, indexing with a loss limit amounts to partial indexing. For assets with nominal losses, it does nothing at all." (Burman, 1999:123).
Indexing with a loss limit would be more disadvantageous for risky and income-producing assets.

Equity
Ensuring the tax system is equitable is of fundamental concern to Government. If only capital gains are indexed, investors receiving capital gain income would be advantaged over investors receiving interest income or other fixed income streams (e.g. through purchased annuities). An investor receiving income in the form of capital gains would only be taxed on the real growth in the asset’s value, while a person that placed his/her funds in a fixed deposit would be taxed on the full – nominal – value of the interest received. This would certainly compromise the principle of horizontal equity.

It is generally accepted that capital gain income accrues disproportionately to wealthier members of society, who would therefore also benefit more than others do if the capital gain income they receive is fully indexed, while other capital income is not. This would be an unacceptable infringement on the vertical equity of the tax regime. Clearly, the equity of the tax system would be significantly enhanced if the tax on all forms of income were adjusted for inflation – i.e. capital gains, interest income (and expenses), depreciation allowances and inventories. However, only a few countries – mainly those that have experienced chronic inflation – have introduced such comprehensive inflation adjustment.

Administrative complexity
Indexing for inflation is conceptually simple – when an asset is sold, the capital gain is the difference between the selling price and the acquisition price, adjusted to take account of inflation between the acquisition date and the selling date. This is so, when considering a basic asset that is bought and sold. However, the administrative and compliance burden increases rapidly if one considers more complex cases.

If an asset is built up over a period of time, e.g. the acquisition of unit trusts or shares on a monthly basis, in an unindexed capital gains tax system, the base cost can be pooled (e.g. using a weighted average cost of acquisition) when the asset (or only a portion of the asset) is sold. If the capital gains tax system were indexed, one would need to keep detailed records in respect of each acquisition of assets. When a sale takes place, calculating the real gain is particularly complex, especially if only a portion of the asset is sold. For instance, one would need complicated rules to match the assets sold with those bought over a period of time. This would dramatically increase the administrative burden on the revenue authority, as well as the record-keeping and other compliance costs for taxpayers.

The compliance burden on taxpayers is captured by Burman (1999:128):
"Indexing would require extensive calculations for an asset that is improved over time, such as apartment building or a business. Every improvement would have to be treated as a separate asset with a separate indexing adjustment."
The more accurate the indexation system the more significant are these issues. If monthly index adjustment tables are adopted, as was the case in Australia and the United Kingdom, the costs of all assets built or acquired over a period longer than a month must be separated into monthly sub-assets, with accurate cut-off procedures to ensure expenses are allocated to the correct sub-assets.

Taking a factory built and equipped over the period of a year as an example, the explosion in the number of sub-assets to be accounted for is apparent. Taking the acquisition of units in a single unit trust as a worked example, a person with a monthly debit order, reinvesting the biannual dividends received, over a period of 15 years would have 210 separate base costs to track. While the compliance burden of tracking and adjusting these base costs may be placed on the unit trust manager, this merely shifts the burden and will leave it with a taxpayer wishing to keep records to monitor the performance of his or her investment.

The matter does not stop here. Each sub-asset will have an associated original cost that will be used for the purposes of normal tax and an indexed cost for the purposes of capital gains tax, resulting in a further doubling of the costs to be accounted for.

Australia and the United Kingdom’s decision to move away from indexation was also influenced by a desire to withdraw some of the complicating features in their legislations that emanated from indexation.
"A report by the Department of the Treasury, Tax Reform for Fairness, Simplicity, and Economic Growth (1984) called for indexing of depreciation deductions. However, this option was not included in the final tax reform bill. The main objection was apparently that depreciation was thought to be too complex." (Congress of the United States: Congressional Budget Office, August 1990: 16, footnote 2).
Finally, if indexing for inflation is implemented, a double adjustment for inflation may arise in a business context. This follows from the fact that interest paid on a loan to acquire an asset will contain a component to compensate the lender for inflation. The borrower will be able to claim both this component and the indexation adjustment as deductions. As a result a mechanism to link loans and assets in such cases will have to be sought in order to disallow the indexation adjustment. The enforcement of any such mechanism is likely to be administratively difficult in a large enterprise with substantial cash flows.

What index to use
Another complicating factor in adjusting the capital gains tax for inflation is deciding the most appropriate index to use:
"In fact, there is no exact cost-of-living adjustment, so even pure indexing would correct only imperfectly for the effects of inflation." (Burman, 1999: 122).
Various indexing proposals have been advanced. Some argue that using a general inflation index for all assets would lead to arbitrariness, unless the inflation-related increase in the base cost of the asset on which the capital gain is realised is the same as the index chosen. This is not likely to be the case for all assets. Using other basic methods (such as applying an approximate adjustment factor) would exacerbate this arbitrariness, leading to more administrative complexity and compliance costs, without realising the advantages of accurate indexation.

The proper indexing factor for capital gains is a measure of general price inflation, like the consumer price index (CPI). It is not correct to index using a measure specific to a particular asset or asset class. The reason indexation is done is to account for the loss in purchasing power of the original investment due to inflation. That loss in purchasing power is a real capital loss. It really has nothing to do with the replacement cost of the machine or building.
Macroeconomic – inflationary expectations
A central part of Government macroeconomic policy has been to reduce and stabilise inflation in South Africa. Considerable progress has been made in this regard, as outlined in figure 1. The "high-inflation" of the 1980’s has been considerably reduced, with inflation in 1999 reported at 5,5 per cent. In the 2000 Budget, Government indicated that the South African Reserve Bank would implement inflation targeting, aiming for an inflation level of 6% - 3% by 2002. In context of lower inflation rates, the additional complexity associated with inflation indexation is not warranted.

Figure 1: SA inflation - 1980 - 1999


7.3 Inflation and capital gains tax
7.3.1 Introduction
The proposed capital gains tax regime for South Africa does not include explicit indexation, i.e. the acquisition cost is not adjusted to take account of inflation over the holding period when the asset is disposed of. However, the potential impact of inflation was one of a number of considerations (though not the primary factor) that informed the decisions to have moderate (low) "inclusion rates" of capital gains in taxable income, thereby partially adjusting for inflation.

This section of the submission, reviews the potential impact of capital gains tax, by reviewing the interaction of various inflation rates, real rates of return and holding periods on the effective capital gains tax rate. It also considers the potential benefit accruing to the taxpayer because capital gains taxes are deferred until the gain is realised on disposal of the asset, not when it accrues.

7.3.2 Effective tax rate analysis
Inclusion rate, inflation rate and pre-tax real rate of return effects on the effective tax rate
When comparing an ad hoc inflation adjustment mechanism with a full indexation mechanism, there are two important aspects to be considered. First, using an ad hoc adjustment mechanism or providing no inflation adjustment at all will cause the effective tax rate to deviate from the statutory tax rate. Of interest to taxpayers, are the conditions under which the effective tax rate exceeds the statutory tax rate. The second concern is where taxing nominal capital gains, with exclusion could result in taxation of capital itself, i.e. where the effective tax rate is in excess of 100 per cent – because the real after-tax return is negative.

Tables 4 (for individuals) and 5 (for corporations) consider the effective tax rate that results under different assumptions, i.e. inclusion rate, statutory tax rate, inflation rate, pre-tax real return and asset holding period. The analysis here illustrates that a constant inclusion rate, constant statutory tax rate, lower inflation rate, higher pre-tax real return and longer asset holding period, confer tax benefits on the taxpayer.

In respect of individuals, it is shown in table 4 that assuming an inclusion rate of 25 per cent (as posited in the policy discussion document) and statutory tax rate of 42 per cent, if the inflation rate is 5 per cent and the pre-tax real return range of "2 to 10 per cent", the effective capital gains tax rate is always lower than the statutory tax rate, irrespective of the holding period of the asset. Clearly, if the inflation rate remains within the 6% - 3% range in terms of inflation targeting, the resulting effective tax rate would always remain below the statutory tax rates, irrespective of the holding period of the asset.

When inflation increases from 5 per cent to 10 per cent, all other things equal, from after eight years of asset holding, the effective tax rate would always remain below the statutory tax rates. At 20 per cent inflation rate, a 2 per cent pre-tax real return, all other things equal, would only result in the effective tax rate to be below the statutory tax rate from after 13 years of asset holding period. Clearly, in the current inflation environment, these are not entirely realistic expectations of inflation.

It is clear that ad hoc measures to adjust for inflation, including the proposed method of excluding a (large) portion of the capital gain from taxable income will cause divergence of the effective tax rate from the statutory tax rate. Whether the effective tax rate is greater or lower than the statutory tax rate will depend on many factors, including the pre-tax real rate of return, the inflation rate, length of time the asset is held and the "inclusion rate".

Length of asset holding period
Section 7.3.2 considers the impact of the inclusion rate, the inflation rate, the pre-tax real rate of return and the length of asset holding period on the effective capital gains tax rate. The effective capital gains tax rate is affected by the length of time for which the asset is held. Assuming a constant pre-tax real return, constant inflation and constant inclusion rate, the effective tax rate would fall over time. This suggests that inflation compensation arising from a constantly low inclusion rate would increase with time. Both tables 4 and 5 illustrate this point quantitatively.

Table 5 illustrates that when the inclusion rate is 50 per cent, and the statutory corporate tax rate is 30 per cent, for a constant 5 per cent pre-tax real return, when the inflation is 5 per cent, the effective tax rate is 25 per cent after five years of asset holding. However, when inflation is 10 per cent, the effective tax rate is 33 per cent after five years of asset holding. After ten years of asset holding, the effective tax rate would be 25 per cent. This further illustrates that the effective capital gains tax rate falls over time. In other words, the value of a low inclusion rate as an ad hoc measure for compensating for inflation increases over time.

The value of deferral
It was pointed out in section 7 that the correct way to tax capital gains is to tax real increases in the value of assets in the period in which such increases accrue to the taxpayer. It was further noted that with the exception of Italy no country in the world that has a capital gains tax system levies the tax on this basis – all levy capital gains tax on a realisation basis. In other words, a capital gains tax liability only crystallises when the taxpayer sells an asset.

There is a benefit to the taxpayer from not paying the tax on accrued capital gains, or postponing this payment until the capital gain is realised. This benefit arises because instead of paying money to the fiscus for capital gains tax on accrued capital gains, the taxpayer is able to invest the funds and enjoy the returns therefrom. For example if a taxpayer holds an asset for three years, in a pure capital gains tax system, a tax would be due each year on the capital growth of the asset. Instead, the tax is only due when the asset is sold after three years. If the taxpayer were to invest the capital gains tax he should have paid each year, he would be able to enjoy the return from that investment. The value of the tax deferral would depend on the after-tax real return from the investment made by the taxpayer and the length of time over which that benefit (tax-free interest) is compounded.

Figure 2 illustrates the value of this deferral benefit to the taxpayer, and makes the following assumptions: the taxpayer is an individual liable for tax at a statutory rate of 42 percent, the inclusion rate is 25 per cent, inflation rate is 5 per cent, the real pre-tax return is 5 per cent, different asset holding periods are used (from one year to twenty years), the initial value of the asset is R1000.

Emanating from the above assumptions, consider a taxpayer holding a growth asset for three years and then selling it. At the end of the first year, he would enjoy no tax deferral benefit and a 61.50 per cent low inclusion benefit on effective tax rate. In other words, the effective tax rate would be 61,5 per cent lower than the effective tax rate on other capital income such as interest.

At the end of the second year, he would enjoy a 2.30 per cent deferral tax benefit and 60.06 per cent low inclusion benefit, resulting in his total tax benefit increasing from 61.50 per cent to 62.36 per cent. At the end of the third year, he would enjoy a deferral tax benefit of 4.57 per cent and a 58.61 per cent low inclusion benefit, resulting in his total tax benefit increasing from 62.36 per cent to 63.18 per cent. This procedure is followed to calculate the deferral tax benefit, low inclusion benefit and total tax benefit for longer time periods, as illustrated in figure 2.

It is clear from figure 2 that the value of the total tax benefit (deferral plus low inclusion rate) increases the longer the asset is held for, a clear and measurable benefit to the taxpayer. As to the interaction between the tax advantage due to pure deferral and pure low inclusion rate, figure 2 indicate that the longer the asset holding period, pure deferral tax benefit is incrementally outweighing pure low inclusion tax benefit.

Figure 2: Total tax benefit due to realisation basis of taxation & low inclusion rate.


7.3.3 Summing up
Inflation can exert important influences on the tax system of any country. It is often argued that these are more acute in the capital gains tax system. It is not in dispute that taxing nominal capital gains can result in effective tax rates that diverge, perhaps significantly, from the statutory tax rates. Section 7.3.2 focussed exclusively on the effective tax rates that could arise under the proposed capital gains tax system. It can be submitted that:

i. The effective tax rate depends on the real rate of return, the inflation rate and the period for which the asset is held.
ii. Given the low inclusion rates proposed, only moderate real returns are required at realistic assumed inflation rates to ensure the effective rate of tax is not greater than the statutory rate of tax, irrespective of the holding period.
iii. The effective tax rate falls as the holding period increases (i.e. deferral benefit).
iv. Taxpayers receive significant benefits because capital gains taxes are levied on a realisation basis, which obviously increase over time.
The central question, however, is whether the ad hoc adjustment mechanism proposed provides sufficient compensation for the effects of inflation. It is illustrated that at moderate inflation levels (below 8%) and moderate real rates of return on assets, that the low inclusion rate and deferral benefits provide significant compensation for the effects of inflation.

7.4 COUNTRY SURVEY
7.4.1 Indexation in the Capital Gains Tax
When contemplating the introduction of a new tax, South Africa can learn much from the lessons of other countries. Relying mainly on the International Bureau of Fiscal Documentation (IBFD) information, a survey of 40 countries was conducted, with respect to the application of capital gains tax and whether such capital gains tax systems allow for inflation indexation.

An overview of the status of these countries with respect to inflation indexation is set out in table 6. Only 3 countries, Hong Kong, Nepal and New Zealand, have no capital gains tax. All other countries surveyed taxed capital gains. 10 out of 31 developing countries were found to have some form of inflation indexation in the capital gains tax system. Argentina and Brazil have since removed inflation indexing, as have the United Kingdom, Spain and Australia. When the United Kingdom and Australia removed indexation, the policy proposals were supported by a need to bring the tax system in line with international practice (Australia) and to simplify the tax system, thereby reducing administrative and compliance costs.

Table 6: A status of 40 countries in regard to having CGT and indexing

7.4.2 Inflation experience
The consumer price index (CPI) of 27 of the countries mentioned in Table 6 as well as of SA are shown in Appendix I. By far the majority of countries that index the capital gains tax for inflation have experienced excessive inflation – e.g. the average for Brazil and Argentina between 1970 and 1999 was 612 per cent and 317 per cent, respectively. Clearly, in an inflationary environment such as this the integrity of the tax system is destroyed unless general inflation indexing is introduced. However, of relevance is the fact that now these countries have got inflation under control, they are no longer indexing the capital gains tax system for inflation. It has already been noted that South Africa’s inflation outlook is extremely positive.

7.5 Conclusion
This section has sought to consider whether Government should include explicit inflation indexation in the proposed capital gains tax system that will be introduced in South Africa from 1 April 2001. Arising from the discussion here, it is proposed that it is not necessary to include explicit indexation. This proposal is supported by the following:

i. At moderate levels of inflation and defensible real rates of return, the low inclusion rates proposed and the deferral benefits arising because the tax is levied on a realisation basis more than adequately compensate for the effects of inflation.
ii. Recently, South Africa’s inflation rate has fallen and begun to stabilise at relative low rates. In this context, and considering the explicit inflation targets being implemented by the South African Reserve Bank, the outlook for inflation is also very positive.
iii. While conceptually simple, inflation indexation is difficult to administer and imposes considerable record-keeping and compliance costs on taxpayers. Even simple indexation, such as broad-band adjustments to base cost for assets bought in a particular time period (perhaps including a number of years) would be complex to administer, without providing the full benefits of indexation.
iv. Internationally, countries are moving away from inflation indexation in an attempt to simplify their tax systems.
v. While the lack of inflation indexation in the tax system as a whole can distort investment decisions, at moderate and stable levels of inflation these distortions should not be overstated. Furthermore, if only one aspect of the tax system is indexed, other distortions are exacerbated, such as the incentive to favour growth assets (e.g. equities and unit trusts) over other "income-generating" assets (e.g. bank deposits) and to finance investment with debt (as the nominal interest is deductible for income tax purposes).
vi. There are also important equity and efficiency implications that must be considered if only one part of the tax system is indexed. If capital gains are indexed, taxpayers receiving income in the form of capital gain (which tend to be the wealthier members of society) are favoured over other taxpayers, which receive income that is fully-taxable (irrespective of the inflation rate), e.g. interest or annuity income. This could compromise both the horizontal and vertical equity of the tax system.

In the final analysis, adjusting for inflation is only necessary if anticipated inflation reaches significant levels, e.g. in excess of 20 per cent, over a prolonged period of time. Given South Africa’s inflation outlook, it is not necessary to contemplate indexing the capital gains tax for inflation at this juncture.

8.Concluding comments
This submission has sought to place the envisaged introduction of a capital gains tax in South Africa in an appropriate context. It has been argued that while a capital gains tax is complex to administer, it is a reasonable trade-off for the improvements in equity, anti-avoidance strategies and economic efficiency that will arise from the taxation of capital gains.
References
Botaro, Renato, "Comments on Tax Administration and Inflation," in Improving Tax Administration in Developing Countries, ed. by Richard M. Bird and Milka Casanegra de Jantscher (IMF, Washington, D.C., 1992), pp. 267-70.

Burman, Leonard E, The Labyrinth of Capital Gains Tax Policy: A Guide for the Perplexed, (Brookings Institution Press, Washington, D.C., 1999).

Casanegra de Jaantscher, Milka; Isaiah Coelho; and Arturo Fernandez, "Tax Administration and Inflation," in Improving Tax Administration in Developing Countries, ed. by Richard M. Bird and Milka Casanegra de Jantscher (IMF, Washington, D.C., 1992), pp. 251-66.

Chirinko, R.S. 2000 Tax Policy and Capital Formation under the EMU: Perspectives on German Economic Policy. For presentation at "Investing Today for the World of Tomorrow" Bundesbank Spring Conferemce, April 28 – 29, 2000.

Congress of the United States: Congressional Budget Office, Indexing Capital Gains, August 1990.

Chua, Dale, "Inflation Adjustment," in Tax Policy Handbook, ed. by Parthasarathi Shome (IMF, Washington, D.C., 1995), pp. 142-45.

Department of Finance. 2000. Budget Review 2000. Pretoria: Department of Finance.

Department of Finance and South African Revenue Service. 2000. Guide to Capital Gains Tax. Pretoria: Department of Finance.

Evans, C. 1998. The Australian Capital Gains Tax: Rationale, Review and Reform. Australian Tax Forum, Vol 14, no 3.

Fazzari, Steven M., and Benjamin Herzon, Capital Gains Taxes and Economic Growth: Effects of a Capital Gains Tax Cut on the Investment Behavior of Firms, (The Jerome Levy Economics Institute of Bard College, 1996).

Fazzari, S.M and Herzon, B. 1996. Capital Gains Taxes and Economic Growth. Levy Institute Public Policy Brief, no 25/1996. The Jerome Levy Institute of Bard College: Blithewood, New York.

Fazzari, S.M. 1999. Capital Income Taxes and Economic Performance. Levy Institute Policy Notes, 1999/7. The Jerome Levy Institute of Bard College: Blithewood, New York.

Jenkins, Glenn P; Gangadhar P. Shukla, Public Finance in open economies, (Harvard Institute of International Development, 1999).

Katz Commission. 1995. Third Interim Report of the Commission of Inquiry into Certain Aspects of the Tax Structure of South Africa. Pretoria: Government Printer.
King, John R, "Taxation of Capital Gains," in Tax Policy Handbook, ed. by Parthasarathi Shome (IMF, Washington, D.C., 1995), pp. 155-158.

Mclure, Charles E., jr., "Comments on Tax Administration and Inflation," in Improving Tax Administration in Developing Countries, ed. by Richard M. Bird and Milka Casanegra de Jantscher (IMF, Washington, D.C., 1992), pp. 270-73.

OECD. 1994. Taxation and Household Saving. Paris: OECD.

Oliver, R. 2000. Capital Gains Tax – the New Zealand Case. Paper prepared for the Fraser Institute 2000 Symposium, on Capital Gains Tax [September 15 – 17, 2000; Vancouver Canada]

Owens, J. 1997. Emerging Issues in Tax Reform: The Perspective of an International Bureaucrat. Tax Analysts.

Premchand, A, Government Budgeting and Expenditure Controls: Theory and Practice, (IMF, Washington, D.C., 1983), pp. 246-56

Shome, Parthasarathi, ed., Tax Policy Handbook, (IMF, Washington, D.C., 1995), pp. -21

South Africa Foundation. 2000. Capital Gains Tax. Occasional Paper no 2/2000. SA Foundation: Johannesburg.

Stiglitz, J.E. 1988. Economics of the Public Sector, 2^nd edition. New York: Norton.

Thirsk, Wayne, ed., Tax Reform in Developing Countries, (The World Bank, Washington, D.C, 1997), pp. 1-32

Thuronyi, Victor, "Adjusting Taxes for Inflation," in Tax Law Design and Drafting, Volume 1, ed. by Victor Thuronyi (IMF, Washington, D.C., 1996), pp. 434-87.