SAJEMS NS Vol 7 (2004) No 2 387

An Overview of Commodity Tax Reform in 
Southern Africa 
________________________________________________________________ 
 
Z Robinson 
 
Department of Economics, University of Pretoria 
________________________________________________________________ 
 
ABSTRACT 
 
Various studies have emphasised the trade and/or revenue implications of free 
trade.  The purpose of this study is to investigate future tax implications of 
further economic integration.  Additional considerations are whether tax 
competition can become an issue and whether it can be used to the benefit of all 
SADC members. An integrated approach was done of what is needed in terms 
of commodity taxation to reach a workable long-term solution.  This article 
analyses experiences in the developing world with reference to lessons learned 
from developed regions.  The first section provides a theoretical background, 
analysing the meaning of commodity tax.  The second section emphasises the 
importance of fiscal decentralisation in federations and the SADC and the third 
investigates the character of and changes in commodity taxation that could 
occur in the future.   

JEL H20 
 
1 INTRODUCTION 
 
The World Bank (1991) has expressed concern about tax competition and 
harmonisation in developing countries.  This specifically includes competitive 
responses from other countries to incentives (tax holidays, accelerated 
depreciation, tax credits, and favourable resource royalties designed to attract 
foreign capital) used in developing countries.  Unexplained phenomena relate to 
the effectiveness of these incentives to attract foreign capital and trade as well 
as the side effects of these incentives, e.g. tax discrimination, to induce or limit 
growth.  The need for developing countries to harmonise taxes among 
themselves through multilateral and bilateral negotiations is therefore, again 
emphasised.  Apart from finding sensible explanations for problematic tax 
practices, a continuous attempt is also made to find long-term workable 
solutions.  The emphasis falls on Southern Africa, that is, the Southern African 
Development Community (SADC).  
 
An African renewal has been postponed for many years.  With the advent of the 



SAJEMS NS Vol 7 (2004) No 2 388 

new millennium, President Mbeki of South Africa referred to this renewal as an 
“African Renaissance”.  The Millenium Partnership for the African Recovery 
Programme (MAP, 2001) led by Mbeki, Obasanjo of Nigeria and Boutefilka of 
Algeria, and the OMEGA Plan (OMEGA, 2001) proposed by President Wade of 
Senegal, can therefore be seen as a welcome change to previous efforts.  At this 
special summit meeting of the heads of state and government in March 2001, 
the need to make provision for a new streamlined structure for the SADC in 
order to speed up its economic integration and cope better with the crisis in the 
region (e.g. the war in the DRC), was also recognised.  Some of the proposed 
changes sought to create legal structures that would adjudicate issues of trade 
and economics that would not suit the interests of individual members.  The aim 
of the reform process is to address the fears of smaller economies and to 
accommodate their aspirations and interests.  The World Bank (2000: 34) points 
out that both “winners and losers (and more of the former than the latter)” 
should be recognised as the first step, although the persuasion of the winners to 
forgo some of their gains to compensate “influential losers who could otherwise 
stymie the process” should form part of trade and investment reform.     
 
Although the SADC has gone beyond the initial stages of an FTA, tax 
competition and coordination problems as such have not yet surfaced.  The 
possibility of a higher degree of economic integration may yet expose these in 
future.  At the same time the formulation of macroeconomic convergence 
criteria for this region has become essential and this should be dealt with within 
the confinements of certain tax restrictions.  The emphasis in this article is not 
specifically on trade or the revenue implications of free trade because various 
studies have already been conducted in this area. The purpose here is to take the 
lead and investigate the future tax implications of further integration.  An 
additional consideration is whether the current situation can be used to the 
benefit of all SADC members. The idea is to form an integrated approach of 
what is needed in terms of commodity taxation to reach a workable long-term 
solution in terms of macroeconomic stability.   
 
This article attempts to analyse experiences in the developing world with cross-
references to lessons that can be learned from developed regions.  In this 
context, the first section analysis the meaning of commodity tax competition 
and therefore provides a theoretical background.  A short background study of 
fiscal decentralisation is then provided and hence also the institutional character 
of the more “developed” federations in the developing world, namely 
Argentina, Brazil and India in comparison with the SADC.  Deeper integration 
means increased exposure to the consequences of the removal of barriers to 
trade and factor movements and the experiences of selected federations can 
provide useful lessons.  Argentina, however, is a unique case study, and is 
included because of its successes with institutional reform during the 1980s and 



SAJEMS NS Vol 7 (2004) No 2 389

1990s.  On the downside, it is also included to show the effects of fiscal 
insustainability on macroeconomic stability1.  The second section emphasises 
the importance of fiscal decentralisation in federations and the SADC.  The 
third section emphasises the current character of and changes in commodity 
taxation that could occur in the future.  The last section attempts to give an 
overview with some conclusions and recommendations.   
 
If not otherwise mentioned, data and statistical resources utilised in the article 
are mainly from the Finance and Investment Sector Coordinating Unit (FISCU), 
the International Monetary Fund (IMF), the National Treasury of South Africa 
(NTSA), PriceWaterhouseCoopers (PwC), and the World Bank (WB).  Various 
problems arise when one tries to analyse data from developing countries.  In the 
case of African studies, Burkett, Humblet and Putterman (1999) state that the 
main problem is the unreliable nature of the available data.  Different variables 
may be recorded for the same observation in different editions of the same 
source.  The IMF (1998) warns that the results obtained by using data should be 
interpreted with caution.  Before proceeding, it is therefore important to shed 
some light on the international status of developing countries in terms of 
taxation.  
 
 
2 COMMODITY TAXES AGAINST THE BACKDROP OF TAX 

COMPETITION  
 
Commodity taxation is investigated within a tax competition context by various 
authors. This type of analysis is especially applicable within a process of 
economic integration, e.g. in a common market or economic union, and a 
federation.  Different commodity tax rates across borders create distortions that 
in turn induce spillovers or externalities such as cross-border shopping.  The 
most familiar types of commodity taxes are the single-stage retail or general 
sales tax (RST/GST) and the multi-stage or broad-based value-added tax 
(VAT).  The main difference between the two entails different methods of 
collection, with RST on a suspensive system and VAT on a repayment system; 
and the tax base that is being taxed also differs. 
   
With VAT the onus is always on traders to convince the tax authorities that their 
claims for refunds on their inputs are justified, whereas under RST there are no 
such claims (the tax is levied only once at the final destination or on imports).  
The claims for refunds or the tax liability can be computed via subtraction, 
addition or tax credits (invoice method). Detailed records of purchases as well 
as sales have to be kept under a VAT mechanism but not under RST.  
Administrative difficulties may therefore occur more readily with VAT, but it is 
also normally implemented to curb tax evasion and corruption.  VAT is also 



SAJEMS NS Vol 7 (2004) No 2 390 

introduced for minimising “tax-on-tax” for which RST/GST is criticised.  
 
2.1 Mintz and Tulkens (1986) 
 
Mintz and Tulkens (1986: 135) were the first to investigate commodity tax 
competition between independent fiscal authorities.  A two-region economy 
(high-tax and low-tax) where an origin-based commodity tax2 is levied by each 
region, is investigated.  The tax is levied on a private good to finance a local 
public service.  The Nash equilibrium of these tax rates is analysed whilst all 
other private goods are untaxed.  A single region’s market and fiscal decisions 
as functions of the region’s characteristics as well as of its environment, is 
investigated.   
 
The analysis is extended in order to consider simultaneous decisions made by 
the two regions.  A so-called regional market equilibrium (RME) and 
interregional market equilibrium (NCFE) are included in the model.  In these 
two cases, the equilibrium is fully efficient.  The main reason for this is that 
transport costs are so high that no cross-border shopping occurs, either in 
equilibrium or in response to small tax changes.  In these cases, none of the 
interregional externalities described previously appear.  Wilson (1999) argues 
that it is difficult to describe these cases as “tax competition”, because 
governments are not really competing over the tax base.  Not all theorists on tax 
competition, however, share this view.  
 
There is a two-person game and the players are local governments.  The 
strategies are local taxes and expenditure levels, and the payoffs are the regional 
welfare function.  Nash equilibrium is established in this two-person game in 
which there is collusion, i.e. a non-cooperative or competitive situation.  This is 
referred to as a non-cooperative fiscal equilibrium (NCFE).  A NCFE amongst 
two regions that choose optimal tax rates and public services production may 
not always exist due to a significant change in the fiscal or tax reaction 
functions.  This means that a switch from one type of regime to another could 
occur.  Differences in the regions’ government size, as well as tax levels can 
therefore arise from strategic behaviour and not only from differences in tastes 
and endowments.  In the absence of interregional public service spillovers, the 
inefficiency of a NCFE thus arises from two types of externalities, viz.:  
(a) Negative private consumption effects (terms-of-trade effects) that occur 

when an increase in a region’s tax affects the private good purchases of 
the other jurisdiction’s residents; and  

(b) Positive public consumption effects that occurs when an increase in one 
region’s tax, increases the tax base of the other region (see also 
Bucovetsky 1995: 362). 



SAJEMS NS Vol 7 (2004) No 2 391

Emphasis is placed on the fact that tax competition is inefficient under the 
origin (source) principle in both regions and that cooperative policy measures 
may become essential in improving the outcome of the NCFE, in short Nash 
equilibrium.  
 
2.2 Kanbur and Keen (1993) 
 
In Kanbur and Keen’s spatial model of cross-border shopping (1993: 877), it is 
argued that unrestricted tax competition (open borders) can take place between 
small and large regions (countries).  The following assumptions are utilised: 
(a) There is a partial-equilibrium model of two countries (home and host) 

and a single taxed good; 
(b) The population is distributed uniformly in each country, but the two 

populations may differ in size; 
(c) Commodity taxes are levied on a destination basis3, and there are no 

barriers to the entry or exit points of new stores; and   
(d) The individual has two decisions to make when buying a commodity, 

viz. to buy in the home country or to travel to the host country with 
transportation costs involved. 

 
A pay-off matrix can be utilised to show the results of the different game 
situations.  In this case, the matrix (Table 1) describes unrestricted tax 
competition as a “prisoners’ dilemma”.  Prisoners’ dilemma is a famous case in 
game theory literature.  Although the analysis is given in terms of commodities 
(cross-border shopping), the same analysis can be applied to mobile capital 
(Hallerberg, 1996).  
 
Table 1 Prisoner’s dilemma facing two regions on tax policy 
 

Region 
b 

Region a 

 
Confess (compete in 

taxes) 

 
Deny (no tax 
competition) 

 
Compete in taxes 

 
3,31

 
6,0 

 
No tax 
competition 

 
0,6 

 
2,2 

Note: 1  These values (x, y) represent pay-offs in terms of ordinal utility 
between A and B (the higher the values, the better the pay-offs).  

 
In Table 1 it is shown that Nash-equilibrium (3,3) is reached where both regions 
A and B both compete in taxes.   When small and large regions compete in 



SAJEMS NS Vol 7 (2004) No 2 392 

taxes, both behave in a Nash manner.  This means that each region chooses its 
own tax rate to maximise its tax revenue while assuming a fixed tax rate by the 
other region, bearing in mind the impact on cross-border shopping.  In this 
equilibrium situation the small region (size relating to the number of residents) 
undercuts the large region because the small region’s tax rate (t) is below the 
large region’s tax rate (T), i.e. t<T.  This point, however, is Pareto-inefficient.  
When neither region competes in taxes, i.e. tax competition is restricted (by 
closed borders), joint tax revenues are reduced or minimal, with larger regions 
suffering a revenue loss (if the difference in size is sufficiently great) and the 
small regions normally gaining revenue.     
 
Cnossen (1990: 476) argues that the potential revenue loss with cross-border 
shopping may be particularly injurious to smaller regions because these regions 
are normally rate-takers and not rate-setters as is the case in larger regions.  
Smaller regions normally tend to set rates lower to increase the volume of their 
sales.  It is therefore undesirable to set a uniform tax rate somewhere between t 
and T because this will always harm small regions, relative to the Nash-
equilibrium.  It will, however, be beneficial to the large country (relative to the 
Nash equilibrium or unrestricted tax competition) if harmonisation takes place 
at rate T (rate-setter); but harmful if harmonisation takes place at rate t (rate-
taker).  Setting a minimum tax rate somewhere between t and T, will lead to the 
small region setting the minimum rate, still undercutting the large region.  In 
this case both regions will set their tax rates higher at point (2,2) and, therefore, 
revenue will increase in both the large and small regions.   This point is thus 
Pareto-efficient because there is no other strategy choice that makes both 
players better off. Tax competition is therefore inefficient.  Kanbur and Keen 
(1993: 889) offer two criteria to determine the optimality of coordination, viz. 
Pareto-efficiency and joint product or revenue maximisation.  In the latter 
instance it is possible to make compensating transfers between the revenue-
losing and revenue-gaining regions. 
 
In the next section a short background study of fiscal decentralisation is 
provided and hence also the institutional character of the more “developed” 
federations in the developing world, namely Argentina, Brazil and India in 
comparison with the SADC.    
 
 
3 FISCAL DECENTRALISATION WITHIN FEDERATIONS AND 

THE SADC 
 
Argentina and Brazil, two of the largest federations and democracies in Latin 
America, have been chosen as examples.  Together with sub-Saharan Africa 
(SSA), Latin America has the highest number of regional groupings in the 



SAJEMS NS Vol 7 (2004) No 2 393

developing world, often with overlapping membership and objectives ranging 
from limited cooperation in specific areas to full-fledged economic integration.  
Furthermore, Latin America has had significant tax reforms since the 1980s 
with substantial growth in FDI inflows during the 1990s as well as 
macroeconomic instabilities.    
 
India, the largest democracy in the world, has a long history in federal finance 
and has been chosen as the representative of the Asian region.  South, East and 
South-East Asia are currently attracting most of the FDI inflows to developing 
countries worldwide (UNCTAD, 2000) with China at the forefront.  The region 
becomes even more interesting if one takes the effects of the financial crisis of 
1997 into account, but at the same time realising the advantages of global 
production networks and attracting most of the parts and components trade 
worldwide (see World Bank, 2000: 66).  Africa, which is still struggling to 
become a major market player, can therefore learn from these and other 
experiences which are also unique in terms of sub-national commodity taxation 
in Argentina, Brazil and India.  In the following sections, fiscal decentralisation 
which is prominent in the three federations, will be discussed and where 
possible compared with the SADC.  Although fiscal decentralisation is not yet 
relevant to the SADC which still has a long way to go before being transformed 
into a common market or economic union, it lays a foundation and provides a 
summary of what these economies could become together or individually.    
 
Figure 1 provides a summary of the fiscal decentralisation features in the 
federations in question and those members of the SADC, which possess actual 
components of decentralisation. The table provides only a glimpse at the degree 
of fiscal decentralisation and further discussions should shed some light on the 
different countries’ tax and expenditure legislation. 
 
Latin America has had a long tradition of centralisation, which dates back to the 
period of colonial administration.  After the independence movement, 
centralised fiscal structures remained in place, partly because of colonial 
inheritance, and partly because of the need for countries to keep distant 
provinces under one power.  Even today, when compared with the industrialised 
world, the region as a whole remains highly centralised.  While sub-national 
levels of government are responsible for over 35 per cent of total government 
expenditure in industrialised countries, on average, in Latin America the 
corresponding figure is less than 15 per cent (Stein, 1998).  The latter, however, 
does not refer to the federations within Latin America (Fig 1).  Although the 
region remains highly centralised, the tendency towards decentralisation is quite 
strong.   
 



SAJEMS NS Vol 7 (2004) No 2 394 

Figure 1 The composition of sub-national shares in selected federations 
and the SADC  

0

10

20

30

40

50

60

Argentina Brazil India Botswana South Africa ZimbabweCountries

P
er

ce
nt

ag
e 

(%
)

Share of government expenditure 1990 Share of government expenditure 2000

 

0
5

10
15
20
25
30
35
40
45

Argentina Brazil India Botswana South Africa Zimbabwe
Countries

P
er

ce
nt

ag
e 

(%
)

Share of government revenues 1990 Share of government revenues 2000
 

Source:  IMF, 2002; See also Table A.1 
 
In Latin America, revenues were traditionally decentralised before expenditure 
responsibilities, and national governments therefore had to maintain spending 
levels with a smaller resource base resulting in large deficits.  For instance, 
Argentina’s subnational governments’ share of total governments expenditure 
declined from 46,3 per cent in 1990 to 43,9 per cent in 2000 while the opposite 
happened in government revenue.  The national government’s share has 
therefore declined even further in terms of revenues.  Separate tax and spending 
powers have allowed subnational governments to incur only a fraction of the 
political and financial costs of their expenditures, especially when most local 
resources are funded from a common national pool of tax revenues creating a 
commons problem.  Subnational governments have controlled substantial 
resources (e.g. the subnational VAT in Brazil), and in some cases have adopted 
revenues from these resources as their “own”.  Both national government and 
local governments levy limited forms of VAT but the state VAT is by far the 



SAJEMS NS Vol 7 (2004) No 2 395

most important of the three and it may therefore be difficult to find a solution to 
this problem.  Although Brazil has gone a long way in the process of devolving 
revenue sources and expenditure functions to subnational governments and 
granting significant autonomy in policy making, local revenue mobilisation has 
hardly been encouraged by the country’s system of intergovernmental transfers.  
A significant vertical fiscal imbalance has arisen in the Latin American region 
typically because more expenditure responsibilities have been assigned to local 
governments with limited own revenues.   
 
The vertical fiscal imbalance has been greater than that in industrialised 
countries with an average of 42 per cent for OECD members compared to 52 
per cent in Latin America.    The vertical imbalance also seems to vary 
significantly between different countries in Latin America, and among 
decentralised countries the difference between Latin America and the OECD 
seems to be even greater.  This suggests that finding a suitable tax base to assign 
to subnational governments is more difficult in the case of developing countries 
(Stein 1998: 105).  The high degree of vertical imbalance in decentralised 
countries in the region creates the possibility of a commons problem, in 
particular when combined with highly discretionary transfer systems, or a large 
degree of borrowing autonomy.  For instance, central bank bailouts to state 
banks that are “too big to fail” have been important in some Brazilian states, 
such as São Paulo and Rio de Janeiro.  These tendencies can also be compared 
with tendencies in Asian economies.       
 
The fiscal systems of East Asian economies have traditionally been highly 
centralised.  In India, states have been granted substantial powers to tax and 
spend but these powers have not been properly implemented.  Central 
government has maintained control over substantial resources and the states 
have had to cope with financing large spending functions with limited 
resources.  However, this unitary behaviour appears to be changing gradually 
with a process of adjustment for states with high deficits and debt.  This trend 
can be partly observed in an increasing tendency in both expenditure and 
revenue shares of subnational governments.  Subnational shares have increased 
in terms of government expenditure (from 51,1 per cent in 1990 to 53,3 per cent 
in 2000) and with a greater but not significant amount in terms of revenues 
(from 33,8 per cent in 1990 to 36,1 per cent in 2000).        
 
An overall summary of fiscal decentralisation in developing and developed 
countries provides insightful reading.  Subnational expenditures comprise a 
small share of government expenditures (the median for developing countries is 
about 14 per cent), except in industrialised countries (the median is about 34 per 
cent) and large federations such as Argentina, Brazil, Canada, India, Mexico, 
the Russian Federation and the US (the median is approximately 45 per cent).  



SAJEMS NS Vol 7 (2004) No 2 396 

Experience in these countries suggests a few guidelines for decentralisation.  
Firstly, in the case of developing countries, fiscal decentralisation is likely to 
generate imbalances at the subnational level which may lead to a deterioration 
of the fiscal position of the national/central government.  As a result, the growth 
performance of these economies may be negatively affected.  Secondly, mature 
federations and EU members have nevertheless experienced higher subnational 
spending shares for a much longer period than most developing countries 
mentioned, without significant fiscal imbalances at the centre.  In these 
countries, more stringent control of subnational fiscal positions (applicable 
budget constraints with explicit transfers) seem to have prevented the 
deterioration of national and subnational fiscal positions owing to 
decentralisation.   
 
Returning to Figure 1, most SADC members have centralised or unitary 
governments and are characterised in some cases by authoritarian rule and/or 
high military expenditures especially in war-torn countries such as the DRC.  
Several heads of state, for instance, those of Namibia and Uganda4 (with the 
most recent case in Zambia) have also opted to adjust their constitutions in order 
to lengthen their terms of office.  Data are therefore not always available or 
nonexistent for subnational levels.  Furthermore, in a number of countries 
decentralisation has not yet resulted in relinquished control from the centre and 
this is partly related to the quality of governance at different levels.  Ghana, 
Malawi and Zambia have each created local councils, but the national 
government continues to direct almost all subnational spending and 
management decisions.  Similarly, the ruling national party in Tanzania holds 
almost all subnational offices.  Besides Uganda, South Africa is regarded as one 
of the few African countries that is in a process of unification through 
decentralisation. 
       
South Africa is a much larger country than, say, France, Germany, the 
Netherlands and the UK combined, and compares well with these unitary 
countries in terms of revenue decentralisation.  The share of subnational 
revenues remained relatively constant from 1990 to 2000 but increased to 
approximately 10 per cent.  In comparison, a substantial change has occurred in 
terms of government expenditures and according to the new constitution (1994), 
the provinces have received many more responsibilities regarding expenditure.  
South Africa’s national government therefore has a relatively small share in 
expenditures (about 50 per cent) in comparison with the European countries 
mentioned.  This makes sense because the larger a country is, the more local 
governments it is bound to have. The provinces have a share of about 40 per 
cent of general government expenditure but relatively little own revenue, 
creating the need for a large amount of downward funding (unconditional and 
conditional or block grants).  This will probably change in the future with the 



SAJEMS NS Vol 7 (2004) No 2 397

South African constitution making provision for a higher degree of devolution 
in terms of taxes.  The process and methods of implementation (either through 
tax surcharges or tax sharing) is, however, still uncertain and future legislation 
should provide clarity on this issue. 
 
From the discussions on fiscal decentralisation, it should be clear that 
decentralised fiscal systems offer a greater potential for improved 
macroeconomic governance, if managed correctly, than centralised fiscal 
systems.  Mature federations in the developed world are proof of this 
conclusion.  Decentralised fiscal systems require greater clarity in the roles of 
various players (centres of decision making) and transparency in rules that 
govern their interactions, to ensure fair play.   
 
 
4 COMMODITY TAXATION AND SOME OF THE 

IMPLICATIONS FOR TRADE  
 
Although in the short and medium term, the trade effects of the establishment of 
a free trade area (FTA) in Southern Africa are of particular importance for 
SADC members, more emphasis is placed on future issues concerning 
commodity taxation in the region.  The liberalisation of regional trade will have 
definite fiscal effects with short-term loss of tax revenues, for instance those 
from SACU.  Fiscal reform and specifically tax reform in this case, should 
therefore be initiated so that the already vulnerable macroeconomic position of 
SADC members is not exacerbated.  Tax reform in this regard again entails the 
objectives of taxation.  The distortionary influence of taxation on consumption 
(savings) and investment should be minimised (ensuring neutrality) whilst 
administrative costs should be kept as low as possible to ensure effectiveness. 
 
The SADC technical arm spearheading trade negotiations, the Trade 
Negotiating Forum (TNF), has agreed on the various policies that are required 
to underpin the implementation of the Trade Protocol.  It mainly recognises the 
need for harmonisation and also includes a need for macroeconomic stability.  It 
also promotes the idea of a future common market set for after 2006.  Various 
studies have been conducted on the introduction of the SADC FTA and its 
effects on Southern Africa (CREFSA, 1998; Evans, 1997; 1998 & 2000; 
Akinkugbe, 2000; Roberts, 2000).  The conflicting effects of other bilateral 
agreements on the SADC are included in these studies, for instance, the likely 
impact of the FTA between the EU and SA on the remaining member countries 
of the SADC which are neither party nor signatory to the agreements.  At 
present, the other SADC-members belong to a wider regional grouping known 
as the African Caribbean and Pacific (ACP) group to which South Africa only 
holds qualified membership because it was not regarded as a typical less-



SAJEMS NS Vol 7 (2004) No 2 398 

developed country.  South Africa was therefore excluded from any agreements 
between the EU and the ACP under the Lomé Convention.   
 
In March 1996, the EU mandate of offers to South Africa for the FTA was 
formally tabled and negotiations commenced for the EU-SA FTA.  Several 
rounds of negotiations were completed by the end of March 1999, giving rise to 
the Agreement on Trade, Development and Cooperation between the European 
Community and the Republic of South Africa of 1999.  The agreement was 
finally signed on 11 October 1999 in Pretoria and the effective date of 
implementation fixed for 1 January 2000.  The salient feature of the agreement 
is that the EU-SA FTA will be established over a transitional period, lasting on 
the South African side for a maximum of 12 years, and on the EU side for a 
maximum of 10 years from the date on which the agreements take effect.  A 
phased elimination of duties is therefore designed in the agreements that will 
eventually lead to free movement of goods, services and capital between the EU 
and SA.  The EU-SA FTA could therefore change the competitive advantage 
that the rest of the SADC members have and South African exports could even 
replace part of the other SADC members’ current exports to the EU.   
 
Akinkugbe (2000: 21) finds that “the implementation of the EU-SA FTA is 
almost parallel with that of the Uruguay Round negotiations (WTO agreements) 
around the world, in the sense that the competitive conditions of the SADC 
members vis-à-vis Europe stand to be fundamentally altered in the next decade 
or so”.  As already mentioned significant structural changes due to 
macroeconomic imbalances are thus necessary in those SADC economies 
planning to diversify the composition of their export trade.  Furthermore, SACU 
members, particularly Swaziland and Lesotho, may lose a sizeable proportion of 
their annual fiscal revenue on the full implementation of the EU-SA FTA, and 
these countries will have to find other ways of diversifying their internal 
revenue bases rather relying on trade taxes.  These effects are also similar to the 
full establishment of the SADC FTA and it is therefore realistic to think with 
the establishment of the SADC FTA with South Africa as one of the members, 
that the EU-SA FTA and the EU-ACP agreements will be incorporated into 
further integration measures to simplify the process. 
 
4.1 Trade within Southern Africa 
 
Observing the trade behaviour of the SADC members identifies some 
interesting characteristics.  The SADC has successfully increased intra-regional 
trade over the past decade.  The trade flows have increased more than tenfold 
since the formation of the SADCC.  The SADC’s performance has clearly 
outweighed that of other regional groupings within Africa, which reinforces the 
notion of the SADC being the most successful integration scheme in Africa 



SAJEMS NS Vol 7 (2004) No 2 399

(World bank, 2002).  In 1997, the World Bank estimated intra-regional SADC 
exports to be over 11 per cent of total exports.  However, many of the SADC 
members’ major trading partners are still outside the community and the African 
continent.  A significant share of the region’s trade is conducted with the 
developed world and the EU in particular.   
 
Over the past two decades, for instance, almost 40 per cent of all SADC exports 
were destined for Western Europe. There is also a strong correlation between 
trade patterns and their colonial ties.  About 13 per cent of all SADC exports 
over the past two decades went to their former colonial rulers.  Also, as already 
pointed out earlier in the discussion, trade levels are low because of the 
composition of their exports.  They often find themselves dependent on the 
export of one single commodity.  For instance, exports from Angola are 
predominantly oil (86 per cent) and for Botswana, diamonds (88 per cent), 
while Malawi’s exports mainly comprise tobacco (76 per cent).  It is also a well- 
known fact that high dependence percentages make economies more vulnerable 
to fluctuations in price and therefore market conditions.      
 
Table 2 South African trade within the SADC (excl. SACU), 1998-2001  
 

Percentage of total SA 
exports 

Percentage of total SA 
imports Country 

1998 2001 1998 2001 
Angola 6,9 7,2 0,5 0,1 
DRC 6,4 3 0,8 0,6 
Malawi 7,7 7,6 21,3 10,9 
Mauritius 6,6 11,7 1,4 2,4 
Mozambique 16,9 24,9 10 12,3 
Seychelles 1,1 1,1 0,4 1,9 
Tanzania 6,6 6,6 1,3 1,3 
Zambia 13,5 21,1 9,9 7,7 
Zimbabwe 34,4 23,3 54,5 64 
SADC  11,1 11,8 11,1 11,2 

Source:  Department of Trade and Industry (2001) 
 
Besides Western Europe, South Africa is the largest trading partner of other 
SADC members. Trade between South Africa and the rest of the SADC 
countries increased dramatically during the period 1990 to 2001 (see Table 2).  
Imports from the SADC increased from less than 1 per cent of total imports in 
1990 to more than 11 per cent of total imports in 2001.  The increase of South 
Africa’s exports to the region was even greater – from 5 per cent of total exports 
in 1990 to almost 12 per cent of total exports in 2001.  South African exports 



SAJEMS NS Vol 7 (2004) No 2 400 

are concentrated in the value-added sectors such as minerals and base metals, 
chemicals, machinery, transport equipment and food and beverages.   
 
The rapid increase in trade with SADC countries during the 1990s shows that 
South Africa enjoys a relative advantage in accessing these markets, mainly by 
virtue of its geographical proximity and South African businesspeoples’ 
networks with their counterparts in the different SADC countries.  The same, 
for instance, applies to Germany which enjoys a trade advantage in the EU 
towards neighbouring CEECs.  Cross-border effects such as trade spillovers and 
cross-border shopping is therefore evident between South Africa and the other 
SADC members, rather than between the SADC members themselves.      
 
South Africa experienced significant trade reforms and liberalisation because of 
its economic isolation prior to 1994.  The import tariff rate (weighted average) 
declined significantly from 21 per cent in 1994 to 15 per cent in 1998.  Textiles, 
electronic goods and automobiles are considered to be price elastic or sensitive 
and tariffs on these items will be phased out over the longer term.  Trade 
liberalisation and change will continue in line with South Africa’s commitment 
to the WTO (see NTSA, 2003).  Since 1999, South Africa has concluded non-
reciprocal bilateral trade agreements with Malawi, Mozambique, Tanzania, 
Zambia and Zimbabwe which should increase trade even further between South 
Africa and these countries.  Under the EU-SA FTA, South Africa has also 
agreed to remove barriers on 86 per cent of EU imports, while the EU will scrap 
95 per cent of its tariffs on South African goods phased in over a 12-year 
period.   
 
At this stage, the average tariff rate in the countries of Southern Africa is about 
12 per cent and this compares well with federations such as Argentina and 
Brazil (13,5 per cent).  These rates are still uncompetitive compared with 
developed regions such as Japan (2,3 per cent) and the US (2,8 per cent) and 
Europe where the average is 2,7 per cent (WEF, 2003).  South Africa has, 
however, established foreign trade relations but always attempts to improve on 
its trade performances by also improving small, micro and medium enterprises 
initiated though the Department of Trade and Industry’s Ntsika project.     
 
South African manufactured goods account for about 70 per cent of exports to 
Africa, where these goods have been successful because of competitive prices, 
shorter supply routes and a sound understanding of the African market.  South 
Africa is, however, still largely reliant on the export of primary and intermediate 
commodities to developed countries despite attempts to diversify its export 
base.  South Africa’s principal exports are gold (40 per cent of total exports), 
platinum, diamonds, coal, food, wine and manufactured products.  Imports 
mainly comprise capital goods, raw materials, semi-manufactured goods and 



SAJEMS NS Vol 7 (2004) No 2 401

consumer commodities, and originate primarily from Germany, Japan, the UK 
and the US.  By February 2003, the main imports together with Switzerland and 
the BLNS countries, were machinery (30 per cent), chemicals (11 per cent) and 
minerals (8 per cent).  Germany, Japan, the UK and the US are also major 
export markets for South Africa.   
 
From the above-mentioned discussion, it should become clear that fiscal 
adjustment may not be as significant in South Africa as in other SADC 
countries where customs and import duties are still important in comparison 
with other revenue sources.  Also, the notion that the integration among African 
states may be sub-optimal to the integration between African states and higher-
income regions (e.g. the EU).  Furthermore, “regional trade agreements between 
Sub-Saharan African (SSA) countries could be of limited value and could even 
lead to trade diversion and a divergence of per capita income amongst member 
states” (Naude & Krugell, 2001: 502).  This also means that for the SADC it 
would be more beneficial to integrate with the EU than with South Africa, and 
vice versa for South Africa.  Further integration and fiscal adjustment measures 
should therefore rather be implemented as a cooperative mechanism between 
the members to ensure that tax revenues are equally distributed on some basis 
such as the population size.     
 
4.2 Commodity tax coordination and reform 
 
Figure 2 summarises the structure of the domestic tax systems of the federations 
and the SADC. For the SADC members there are still significant disparities.  As 
already mentioned, Malawi, Mozambique, Zambia and Zimbabwe are countries 
that suffer the most from an FTA and/or customs union (Evans, 2000).   
 
In general, tax systems are known to be non-neutral, that is, they cause 
distortions in terms of the allocation, especially in goods and factor markets in 
developing regions.  Although estimates are unreliable, that is, the excess 
burden of taxation is significant in these countries various reforms have been 
introduced mainly to broadening the tax base with a simultaneous reduction in 
tax rates.  In the SADC, members would either have to broaden the tax base or 
increase tax rates to compensate for losses incurred by the planned FTA and/or 
customs union.  Some of these members have already started significant 
adjustments, especially in the field of commodity taxation, because capital 
income taxation does not allow much room to increase rates (see Fig 2) or 
broaden the base.  Marginal income tax rates are already high in comparison 
with international standards and increases in these rates are likely to distort 
employment, savings and investment even more, with a likely increase in tax 
evasion. 
 



SAJEMS NS Vol 7 (2004) No 2 402 

Figure 2 Commodity tax revenue, 2000 
 

0

10

20

30

40

50

60

70

Argentina Brazil India (central) South Africa SADC
Countries

%
 o

f t
ot

al
 ta

xa
tio

n

Sales Taxes Excise Taxes Commodity Tax Capital income Tax

Source:  IMF, 2002; See also Table A.2   
 
In terms of the tax base, the average share of capital income taxes already 
carries a significant weight (31,3 per cent) and leaves some room for adjustment 
compared with the EU average (41,6 per cent).  The worldwide tendency has, 
however, been to rather keep the corporate tax burden as low as possible in 
order to attract more active or fixed capital and even to zero-rate some 
withholding taxes on foreign interest income.  The belief is that capital income 
taxes, specifically those on interest income, are “disappearing” because the 
implementation of the residence principle on interest income has become ultra 
burdensome with cooperation between governments not always forthcoming. 
The SADC member governments also realise the need for the attraction of 
capital and this can be observed from the numerous tax incentives provided in 
terms of export processing zones (EPZ) and foreign investors.  Accepting that 
the SADC governments have already decided on the relevant balances between 
capital income and commodity taxation, the most significant reform measures 
and fiscal adjustment could occur in terms of commodity taxation in the SADC.  
Capital income taxation, however, is not flawless and there is room for 
improvement.   
 
The average share of sales taxes (14,6 per cent) and excise taxes (10,1 per cent) 
in the SADC are still relatively low (Fig 2) compared with the EU average of 
27,4 per cent for VAT and 18,9 per cent for excise duties.  In terms of the 
federations, Argentina’s subnational authorities’ tax revenue as a share of total 
revenue is about 90,8 per cent, Brazil’s 77,4 per cent and India’s 83,4 per cent 
for 2000. Sales taxes (VAT) as a percentage of total taxation in Argentina are 
27,5 per cent, Brazil 8,5 per cent and India only 0,1 per cent.  The same shares 



SAJEMS NS Vol 7 (2004) No 2 403

in terms of excise taxation are 13,8 per cent in Argentina, 7,8 per cent in Brazil 
and 25,7 per cent in India.  Although India’s subnational share seems high, the 
central government still tends to undermine these authorities’ rights and 
exercise to much control over these resources.  The commodity tax shares of 
subnational levels are also much higher than those in mature federations such as 
Canada and the US.   
 
Taxing powers have therefore been significantly devolved in Argentina, Brazil 
and India, and in some cases even to the detriment of the national governments.  
Although this trend of decentralisation is not observable within the SADC (only 
in South Africa), experiences from these countries can serve future prospects of 
SADC members as the region becomes more integrated towards a common 
market and possibly an economic union.  The tax reform experiences of Latin 
American countries such as Argentina and Brazil, which have experienced 
significant tax reforms since 1980, can also provide important insights for future 
reform in the SADC. Latin American governments realised the importance of 
tax policy, and tax reform therefore became an integral part of wide-ranging 
economic reforms in the region.   
 
Latin American governments found that tax policy was an instrument that was 
relatively easier to wield than politically difficult expenditure cuts, and its 
effects were more immediate, and in the short and medium term at least, it was 
more directly measurable than those of other economic policies.  As Latin 
American economies became more integrated with the rest of the world, tax 
systems could no longer be viewed in isolation.  The growth of emerging 
financial markets and the surge in direct investment and more open trade and 
payments regimes gave impetus to the reform movement.  Tax competition 
became a prominent issue and governments realised that they had to reduce or 
eliminate taxes that raised business costs and domestic firms in positions where 
higher rates applied at a disadvantage in world markets.   
 
At the beginning of the 1980s, most Latin American tax systems were complex 
and cumbersome, loaded with hundreds of revenue agencies with little revenue 
being collected.  Consumption and production suffered because of multiple rates 
and were weakly administered (Shome, 1995).  These taxes were also 
insufficient because of “cascading”, which taxed not only the value of 
production but also taxes paid at earlier stages of production since they were 
generally levied at the manufacturing stage (such as those now present among 
some of the SADC members), rather than the retail stage, they hampered 
competition and added to production costs.    
 
Tax reform in Latin America has been implemented in various ways.  In 
Argentina, for instance, tax reform led to a radical redesign of the entire system, 



SAJEMS NS Vol 7 (2004) No 2 404 

whereas in Columbia, it was carried out as a series of steps over a number of 
years.  Reform-induced countries simplified their tax systems focusing on 
income taxation in the early years of the reform process and, increasingly, the 
taxation of production and consumption in later years.  As the economies of the 
region matured and were integrated with the rest of the world, attention was also 
focused on the fine-tuning of particular aspects of the tax system with 
international ramifications such as the exchange of information, foreign tax 
credits and transfer pricing.   
 
In terms of commodity taxation, VAT was an important part of the reform effort 
between the early 1980s and 1994, and the number of countries with a VAT 
doubled from 10 to 20.  In the early 1980s, some countries either had a 
rudimentary VAT up to the manufacturing-importing stage, or a production-
type (origin-based) VAT which disallowed credit for capital goods purchases.  
In the second half of the decade, these countries began to reform their VATs, by 
reducing the number of rates (Bolivia, Chile, Columbia and Mexico) and 
expanding the base by reducing exemptions and raising coverage, particularly of 
services (Argentina, Bolivia, Chile, Colombia and Mexico).  Furthermore, some 
countries, notably Argentina, Chile, Colombia and Mexico converted to 
consumption-type (destination-based) VAT and improved their tax 
administrations.   
 
Countries that achieved a large increase in their tax-to-GDP ratios, such as 
Argentina, Bolivia and Colombia often did so through VAT.  As VAT revenue 
rose, countries relied less on excises, taxing only a few items such as beverages, 
tobacco, petroleum products and automobiles instead of a broad range of goods 
and services.  All of the major countries in Latin America have also done away 
with export duties and most have reformed import tariffs with the dispersion of 
these rates being reduced and tariff levels significantly decreased.  The Latin 
American experience is far from perfect, especially with regard to subnational 
commodity taxation as in Brazil.  Fiscal or tax adjustments should therefore 
always directly be handled in line with economic circumstances in a particular 
country.  As mentioned earlier, some SADC members have already started 
fiscal adjustments for the planned SADC free trade area (FTA) and/or customs 
union (CU), specifically in terms of commodity taxation.  In the next section, 
commodity taxation in the SADC is discussed, specifically in terms of 
adjustment and reforms that might be needed in future.    
 
4.2.1 Fiscal adjustment and tax reform 
 
Table 3 gives an indication of the present commodity tax rates (column 2) 
against what is required in terms of the planned SADC FTA and/or future CU 
(column 3), and also excise tax categories.  Namibia has experienced a 



SAJEMS NS Vol 7 (2004) No 2 405

significant adjustment with a switch from a GST of 8 per cent in 1997 to a 
standard VAT of 15 per cent with an increased rate of 30 per cent in 2000 
(column 2, Table 3).  This rate would leave ample room to compensate for any 
losses in terms of customs revenues.  Tanzania also has a VAT system with a 
rate that changed from 10 to 20 per cent from 1997 to 2000, with the DRC close 
on its heels with a standard VAT rate of 18 per cent.  Although these countries 
did not incur significant problems in compensating for any revenue losses of 
tariff reform, the rate is not as important as the quality of the reform involved.  
Keen and Lighthart (1999: 18) point out that, if an underlying tariff reform 
improves production efficiency, replacing the tariffs with domestic 
consumptions taxes (specifically emphasising VAT) will raise welfare in a 
small open economy.  This should therefore also be applicable to those 
countries in which significant tariff reforms are necessary.   
 
Countries such as Malawi, Mozambique, Zambia and Zimbabwe would lose the 
most in terms of the planned FTA and/or customs union (column 3, Table 3) 
and therefore require close scrutiny.  Malawi has a so-called SUR-tax (between 
sales and value-added tax) of 20 per cent.  CREFSA (1998) estimates that this 
rate would have to increase by 5 percentage points to offset revenue losses.  
Mozambique has a standard VAT of 17 per cent which was introduced in 1999 
and as such is already on the road to change.  Zambia has a standard VAT of 
17.5 per cent introduced in 1995.  According to CREFSA (1998), this rate 
would have to increase by 3 percentage points to offset losses.  The VAT in 
Zimbabwe would have to change by 8 percentage points to 23 per cent if the 
present standard rate of 15 per cent on goods and services other than basic 
goods (0 per cent), electricity (5 per cent) and luxury goods and motor vehicles 
(25 per cent) is taken into account.  
 
Mauritius, which has been relatively dependent on trade taxes, has a VAT of 15 
per cent bringing it more in line with the rest of the SADC members.  Besides 
rate increases, the mere fact of switching to VAT in the SADC can already 
improve the buoyancy of these revenue sources and decrease the distortions 
associated with commodity taxation.    
 
Any sales tax that does not extend through the retail sales level can cause 
administrative problems because if the actual sales price is used, competitive 
distortions are created in different channels of distribution.  For instance, when 
manufacturers or wholesalers sell directly to ultimate consumers at retail prices, 
the price should presumably be reduced to the wholesale level.  However, if 
manufacturers sell directly to retailers who assume some manufacturing or 
wholesaling functions, the actual sales price should be increased to equalise the 
situation of such sales with those made through regular channels of distribution.  
This type of tax has, therefore, proven to be ineffective as a revenue-raising 



SAJEMS NS Vol 7 (2004) No 2 406 

instrument in several countries such as Australia (see Messere, 1993), and 
countries such as Angola should also consider changing over to a VAT system.  
South Africa’s experience probably served as a directive to those SADC 
countries that switched over to VAT during the 1990s and more recently.   
 
Table 3 Commodity tax rates and excise categories, 2003  
 

Country Sales tax rates (%) 

VAT-
rates 

(SADC 
fta/cu)

Excise categories 

Argentina1 1 – 3.5 (RST) --- Wide variety (other than exports) at varying rates. 

Brazil1 7 & 25 (ICMS) --- Federal excise tax (IPI) which is similar to a VAT.  

India1 4 (CST) --- Levied on certain types of manufactured goods. 

Angola 10 & 30 (sales tax at manufacturing level) N/A 
Other taxes include taxes on oil 
production, an oil transaction tax 
and a surface tax. 

Botswana 
0 & 10 (sales tax at 
manufacturing and 
import level) 2

10 

Tobacco (15 per cent), alcohol, fuel, 
soft drinks and certain luxury goods 
with specific rates on fuel and 
alcohol. 

DRC 0.25-3 (X), 3-13 (local), 18 (construction), 302 N/A Luxury goods, alcohol and tobacco. 

Lesotho 142 11 

Alcohol, fuel, soft drinks, matches 
and certain luxury goods.  Other 
taxes include sand and stone levies 
and a petrol levy. 

Malawi 17.5 (SUR-tax) 25 Tobacco and alcohol. 

Mauritius 152  12 Tobacco, alcohol and materials (textiles). 
Mozam-

bique 17
2 N/A Tobacco, alcohol, perfume, cosmetics and toiletries. 

Namibia 15 & 30 (luxury goods)2 15 

Tobacco, alcohol, fuel, soft drinks 
and certain luxury goods. Other 
taxes include a sales duty (0–25 per 
cent). 

Seychelles 

Trades tax under the 
Trades Act 1992 on all 
imports and locally 
manufactured goods and 
services. 

N/A --- 



SAJEMS NS Vol 7 (2004) No 2 407

Table 3 continued 

Country Sales tax rates (%) 

VAT-
rates 

(SADC 
fta/cu)

Excise categories 

South 
Africa 14

2 14 
Tobacco, alcohol, fuel, soft drinks 
and certain luxury goods.  Other 
taxes include a fuel levy. 

Swaziland  14 & 25 (luxuries) 
 
13 

Imported local goods such as 
tobacco, alcohol and petroleum 
products.  Other taxes include a 
sugar cane levy, sugar export levy, 
entertainment tax, sports levy and 
fuel tax. 

Tanzania 202 11 
Other taxes include a dairy industry 
levy, entertainment tax and motor 
vehicle registration tax. 

Zambia 0 & 17.52 20 Tobacco, alcohol, fuel and entertainment tax. 

Zimbabwe 0, 5, 10, 15 & 25 23 

Tobacco, alcohol, fuel, soft drinks, 
petroleum products.  Other taxes 
include a tobacco levy and 
automated financial transactions.  

Notes: 1 In Argentina, a VAT of 21 per cent is applicable at federal level.  Each 
of the 24 provinces imposes a quasi-RST (taxation on the gross 
revenue for the sale of goods and services) where most industries and 
exports are exempt from this tax.  In India, the central government 
levies a VAT of 16 per cent (the Union excises).  Besides the latter, a 
special tax, namely the central sales tax (CST) of 4 per cent, is levied 
by exporting states on interstate exports (origin-basis) in all the states 
and the revenues also accrue to the states. Brazil has a subnational 
VAT (7-25 per cent) levied on an origin-basis. 

2 All these countries have a VAT system. 
3 The BLNS countries and South Africa as SACU members maintain 

essentially the same tax base in terms of customs and excise duties  
(Customs and Excise Act with Amendments, 1964; NTSA, 2001). 

Source:  PwC (2004) 
 
The South African experience, like the experience of Latin American countries, 
can also provide useful lessons to the rest of the SADC members that are 
planning to switch over to a VAT system.   The collection of commodity taxes 
through a comprehensive VAT (including retailers) has become standard 
practice worldwide.  It has also become the focus of tax reform efforts in 
developing regions.  In South Africa, the Margo Commission (1987: 345) 



SAJEMS NS Vol 7 (2004) No 2 408 

recommended as an alternative to reducing the GST rate and introducing a 
comprehensive business tax (CBT) or a value-added income tax (VAIT), that 
GST should be abolished and replaced with a VAT credit or invoice-based 
system.  Consequently, VAT was introduced on 30 September 1991.  
 
4.2.2 Base broadening and tax compliance  
 
In accordance with the OECD practice worldwide which is a reasonable 
indicator of “international best practice”, VAT should be on a broad basis 
(including goods and services) and a credit (invoice) system, and the destination 
principle should hold.  Thus far, the administrative ideal of only a few rates has 
been achieved in South Africa with only one standard rate and a zero rate.  The 
rate has also been relatively low and increased from 10 to 14 per cent, which is 
still within the recommendation of the World Bank (1991) of between 10 and 20 
per cent for developing countries.  In 1996, the VAT base was broadened to 
include most fee-based financial services.  However, the VAT system tends to 
be more regressive with a few exemptions and zero-ratings still in place.  The 
Katz Commission (1994: 133) recommended against the further erosion of the 
VAT base through zero-rating or exemptions stating that targeted poverty relief 
and development programmes should rather receive priority instead.  This 
currently is being done in South Africa through the Department of Public Works 
community-based projects.  In addition, higher VAT rates on luxury goods or a 
multiple VAT rate system should be avoided.  The main reason for the latter 
recommendation was that such a system would not reduce regressivity, would 
have high administration and compliance costs and would not have much 
additional revenue potential.   
 
In a VAT credit system, multiple rates, as in the case of zero-rating (say, on 
exports) and exemptions, open up opportunities for fake claims and hence tax 
evasion, but also complicate administration for tax authorities and taxpayers 
alike.  In the end, an optimal system of commodity taxation can be secured only 
if the loss of economic efficiency with VAT is minimised through uniform rates 
or a few rates applied to the broadest possible base.  If this does not happen, a 
compromise will have to be made between administrative costs and equity.  
This case is even further strengthened if a system of income and expenditure 
supports is already in place for the poor. 
 
A broad-base VAT also means that the necessary increase in tax rates is smaller 
than for specific commodity taxes such as excise taxes and that the risk of 
distorting specific markets is correspondingly lower.  The SADC region can 
gain from South Africa’s experience of excise duties.  The World Bank (1991: 
6) recommends setting three or four selective tax rates on luxuries and 
nonessentials, with the rate ascending according to the item’s role in the 



SAJEMS NS Vol 7 (2004) No 2 409

consumption of the rich.  The Katz Commission (1994: 133) recommends that 
the present ad valorem excise duties in South Africa be retained but that the 
possibility of introducing a progressive ad valorem duty on luxury motor 
vehicles should be investigated.  Since 1994, excise taxes on tobacco products 
(for health reasons) have progressively been increased to 50 per cent of the 
retail price.  The BLNS countries already have a common customs (which has 
been phased down in line with WTO regulations) and excise system applicable 
under South African legislation for SACU, and it will become necessary for a 
future SADC FTA (CU) to specifically coordinate excise duties for further 
integration purposes.  The South African government has also initiated a rewrite 
of the Customs and Excise Act of 1964, primarily because its readmission to the 
international arena has shifted the focus from revenue collection to trade 
facilitation and control (NTSA, 2000: 73).   
 
At present the excise categories in the SADC largely correspond and include 
mainly tobacco, alcohol, fuel, soft drinks and certain luxury goods (Table 3).  
The rates, however, still seem to differ considerably.  In this regard, the EU 
experience can serve as an example with common customs and tariffs, and 
common excise duties, which have been set through minimum rates.  The 
SADC countries should, however, be careful not to imitate developed regions’ 
experience in every detail because as already mentioned, the needs of 
developing countries’ may differ significantly from those of developed 
countries.   
 
Another more recent experience of the Zambian government could better serve 
the SADC region. The government undertook a comprehensive review of both 
the tax system and customs duties, with the intention of significantly broadening 
the base of taxation.  A considerable number of exemptions in both taxes and 
customs duties were eliminated and the emphasis shifted from specific 
consumption taxes such as excise duties to sales taxes in the form of VAT 
introduced during 1995.  The Zambian authorities achieved higher revenues (tax 
revenue reached 31.5 per cent of the GDP in 2000), despite significant cuts in 
customs duties and marginal tax rates.  The ultimate objective is that other 
SADC governments would learn from, say, the South African and Zambian 
experience.  It is, however, questionable whether the different SADC 
governments cooperate with their neighbours or whether they have established 
links with one another5.   
 
The Zambian experience reminds one of Argentina’s experience with tax reform 
at the beginning 1990s.  The government implemented radical changes in 
response to successive crises, and the lack of political resolve to enforce tax 
laws progressively eroded the tax structure and administration.  Revenue only 
reached 11 per cent of the GDP in 1989, compared with 14 per cent in 1985.  



SAJEMS NS Vol 7 (2004) No 2 410 

The strategy was thus to improve the quality and quantity of revenue 
mobilisation by eliminating taxes that were easy to collect but inhibited growth 
such as export taxes and taxes on financial transactions, and to concentrate 
instead on a few major taxes such as VAT and on overhauling the tax 
administration.  The strategy was highly successful, and the ratio of tax revenue 
to GDP climbed to 16 per cent in 1993.   
 
Argentina’s VAT went from being the least revenue-productive6 in the world to 
being highly productive (Shome, 1995).  The tax base was broadened and 
evasion sharply cut.  Businesses failing to make timely or correct declarations 
were summarily closed for three days.  In 1990, 700 taxpayers were penalised in 
this way.  In 1992, the number rose to 12 000.  This had a strong impact on 
VAT compliance.  New invoicing requirements and controls were introduced, 
and expanded information on VAT taxpayers helped to improve the collection 
of other taxes by permitting tax inspectors to cross-reference tax data.  Since the 
second half of 1992, the government has focused increasingly on using the tax 
system to improve enterprise competitiveness.  Foreign trade taxes have been 
lowered and in an effort to improve the cost structure of the economy as a 
whole, the federal government has also started encouraging provincial 
governments to reduce or eliminate local taxes that impinge directly on 
enterprise costs.  Although the Argentine VAT system is not flawless, it can 
teach the SADC an important lesson on the topics of base broadening and tax 
evasion.   
 
Although it would appear unnecessary with a tax revenue effort already in full 
swing, a broadening of the tax base in the SADC could minimise upward 
pressure on tax rates, and thus improve the international competitiveness of the 
region and avoid excessive tax-induced distortions.  It could also improve the 
revenue productivity of commodity taxes (GST and VAT), which is still 
relatively low measured in terms of OECD standards (CREFSA, 1998).  
Various countries also still maintain tax incentives concerning sales taxes on 
exports or export-processing zones (EPZs).  For instance, in Angola, Malawi, 
Mauritius, Mozambique, South Africa, Zambia, and Zimbabwe all kinds of 
incentives (VAT exemptions, and customs and excise duty exemptions) on 
materials and inputs used for exporting purposes or within an export-processing 
zone (EPZ), are still included in the GST/VAT system.  Although these 
incentives apply to outward-oriented operations, they can reduce the effective 
VAT rate and work against the objective of a broader and ultimately a 
coordinated base for the region as a whole.  SADC governments have, however, 
increasingly started with concerted efforts in tax compliance, notably 
Mozambique where an independent UK-based private agency has been 
appointed to operate customs revenue, and in South Africa, where the South 
African Revenue Service (SARS), an autonomous agency within the public 



SAJEMS NS Vol 7 (2004) No 2 411

service, has been established.  In both cases the result has been a significant 
increase in revenues collected.   
 
Tanzania and Zambia have also established independent revenue authorities and 
it seems as if this type of reform appears to be gaining ground in the region with 
positive effects for revenue productivity. Zimbabwe had the same kind of spot 
calls in 1997 as those in Argentina in 1992.  Under separate operations, the 
Department of Taxes and the Department of Customs and Excises in Zimbabwe 
checked businesses for tax payments and gained substantial sums of money 
outstanding.  In South Africa, the SARS record speaks for itself.  By the end of 
March 2001, SARS had continued its 6-year track record of surpassing tax 
collection targets (National Treasury of South Africa, 2001).  Most of the 
additional revenue was derived from income tax and it would thus appear that 
tax evasion has been considerably reduced.   
 
Although from experience, the most attractive instrument of commodity 
taxation for fiscal adjustment is a broad-based GST in terms of simpler 
administration and compliance, SADC members generally seem to be switching 
over to a destination-based VAT credit system.  The administrative capacity of 
these countries seems to have a greater reliance on VAT, and this could even 
enlarge in future.  Botswana, for example, has decided to introduce a 
destination-based VAT of 10 per cent in 2002, which will make its sales tax 
base much wider but with fewer exemptions and zero-rated exports.  Lesotho, 
has also implemented a 14 per cent VAT on 1 July 20037.  The VAT system is 
usually known for its broad basis with value being added in each production 
stage (multiple stages).  Higher rates are therefore more tolerable in terms of 
VAT rather than a US-type GST or RST levied at one stage.  The VAT system 
normally entails more administrative costs and it is therefore important that the 
introduction of such a system involves careful planning and that governments 
already have the administrative capacity in place to maintain the system.  An 
effective VAT system, if correctly implemented, can therefore eliminate any 
discrepancies that might exist in terms of bookkeeping for tax purposes and also 
tax evasion as such.   
 
Shalizi and Squire (1988: 7) emphasise that if an “embryonic” or all-inclusive 
commodity tax is in place, the role of this taxation as a source of revenue should 
be expanded.  In the short run, this could be achieved by means of an increase in 
the tax rate with a compensating reduction in tariff rates (as already pointed out 
in the case of the SADC).  In the long run, expansion in the base will allow 
further reductions in rates and cause an increasing amount of revenue to be 
generated from the taxation of domestic activities.  Further measures of fiscal 
adjustment in the SADC to reach the objective of regional growth may therefore 
include a broadening of the commodity tax base past the 



SAJEMS NS Vol 7 (2004) No 2 412 

manufacturing/wholesales level as already discussed, and the broadening of 
capital income tax bases with enhanced compliance and disciplined spending.    
 
4.3 Future prospects   
 
It is clear from the investigation thus far that the SADC governments will have 
to make some adjustments in terms of commodity taxation and that there is 
always room for improvement.  In this Section, future prospects in terms of the 
present system of commodity taxation that exist in the region will be explored.  
The possibility of the adoption of a VAT system for the whole of the region in 
particular will be analysed, especially with a view to further integration efforts 
such as a common market (CM).  Under the present system, it would appear that 
most of the SADC governments will apply a destination-based VAT in the 
future, and that the intention in the short run could be to skip the FTA and 
immediately move to a CU where origin is irrelevant. 
 
4.3.1 Revenue sharing, revenue clearance and the destination principle 
 
The SADC might proceed with a CU because of the difficulty of compliance 
with origin rules in the planned FTA and/or CU.  In a CU, countries are 
normally concerned with avoiding fiscal discrimination (i.e. ensuring tax 
neutrality) which causes differences in competitiveness, and the importance of 
fiscal autonomy or sovereignty is recognised.  Customs duties are collected at 
the point of entry, irrespective of the final destination of the goods within the 
CU.  All of the parties to the CU may therefore collect duties on imports 
destined for other parties.  Unless the customs revenues from such cross-border 
or trans-jurisdictional imports are more or less equal, some mechanism or 
formula for revenue allocation must be devised if one of the parties is not to end 
up subsidising the other.  For instance, in terms of SACU, BNLS imports from 
the rest of the world reach South African harbours first and South African 
imports alone are also much more than those of the BNLS countries.  A CU 
therefore abolishes economic but not fiscal borders, and fiscal equalisation 
measures (e.g. the common revenue pool of SACU) are needed to eliminate 
distortions due to taxation.   
 
In terms of the future SADC FTA/CU, countries such as South Africa may not 
be willing to provide compensatory mechanisms such as those applicable within 
SACU.  The uncertainty exists about South Africa’s ability or willingness to 
continue its compensation programme to current or any future members of an 
enlarged SACU (SADC) as well as the willingness of the other members to stay 
in SACU in the absence of such payments.  As SACU continues to liberalise, 
the new revenue-sharing formula will become impractical but will also have a 
negative impact on smaller economies that have done little to diversify their 



SAJEMS NS Vol 7 (2004) No 2 413

revenue source.  A similar problem arises in respect of other indirect taxes, 
especially the VAT, if the destination principle is to be upheld.  It is often 
argued that this principle is rarely supported in full by tax-incidence analysis, 
and that if one considers the implications of the term “VAT” - a tax on value 
added in production - then its revenues should fall to the tax jurisdiction in 
which production (origin) takes place, rather than the one in which the final 
products are consumed (destination).  The application of the destination 
principle, however, ensures that foreign sales (exports) are tax-exempt and all 
domestically paid taxes are reimbursed.  Taxation therefore takes place in the 
importing country, including taxes on the last exchange with compensation 
corresponding to the various taxes that a similar product of the importing 
country would have paid in the preceding phases. 
 
Normally special provisions such as border tax adjustments still have to be 
made to enable the exporter to be rebated for VAT previously paid in the 
importing country.  The importer can then reclaim VAT after retail sale in the 
importing country.  The destination-based VAT therefore guarantees a greater 
degree of neutrality and uniformity.  For a future SADC CU/CM, revenue 
sharing or rebate schemes can be constructed on some general formula or on 
that of actual revenue clearance.  Whereas the former has the advantage of 
administrative simplicity, it also raises questions of the formula’s derivation and 
its adjustment over time.  In the case of VAT, for instance, the general formula 
normally fails to take into account the composition of trade.  For instance, it 
could be that VAT-exempt items constitute a larger share of Zimbabwe’s 
exports to South Africa than of imports from South Africa.  The formula would 
therefore overestimate the VAT revenue on South Africa’s purchases accruing 
to Zimbabwe, thus underestimating the sum to be rebated to it from South 
Africa.  More detailed formulas, on the other hand, would require frequent 
updating and renegotiation.   
 
A bookkeeping approach or national accounts of revenue clearance on the basis 
of actual payments would make more sense for a future SADC.  This will, 
however, require identification of the ultimate destination and that such 
information be readily available from a computerised customs-clearance system 
operated by the South African customs authorities which are now also 
responsible for customs and excises of SACU.  In this case, it is important to 
add that the SADC Ministers of Finance signed a Memorandum of 
Understanding (MOU) on Cooperation in Taxation in 2002.  The MOU will be 
incorporated into the SADC protocol on Trade and Investment that is due to be 
signed in 2004 (NTSA, 2003).  Further, in terms of VAT, a credit system with 
invoices issued by the sellers (registered traders) to the importers will prevent 
double taxation.  The importer normally uses this invoice to claim a rebate from 
his/her tax authorities for VAT already paid on inputs purchased from the 



SAJEMS NS Vol 7 (2004) No 2 414 

exporting country and thus retrieving the money from the exporter’s exchequer 
or revenue authorities.        
 
In the absence of economic borders, say in a future SADC common market 
(CM), the difficulty of levying VAT on a destination basis (same as interest 
income on a residence basis) effectively means that the origin principle is 
practised.  The origin principle is, however, distortive in terms of production.  
The destination principle (with a revenue-clearance system) will therefore make 
more sense in a developing context, specifically because production (export-
oriented growth) and FDIs are first in line as employment generators.  
Furthermore, in a future SADC, competition could become exceptionally severe 
when tariffs have been eliminated among member countries in terms of a CM.  
Here, each country could promise a wider market to foreign capital than before 
and consensus would have to be reached in terms of tax diversity (competition) 
or uniformity.  The same argument applies once revenue losses in terms of the 
SADC FTA have been recouped through, say, higher VAT rates, and countries 
such as Malawi, Zambia and Zimbabwe will have to re-evaluate their different 
situations to prevent tax competition from driving them out of the market.  The 
experiences of other developing countries can also provide useful lessons in this 
context.     
 
In countries such as India and Argentina, commodity tax reforms have 
proceeded more slowly (only started in 1992) than those in Brazil, perhaps 
because of fear of the same difficulties.  India’s sales tax system is still 
problematic because of its non-harmonised nature with various types of sales 
taxes at central and state level, including a CST levied on an origin basis at state 
level (Table 3).  Most of the Asian economies carry a sales tax or VAT-rate of 
10 per cent and as such the World Economic Forum (2003) classifies sales tax 
or VAT rates in Asian economies as a competitive asset whereas Latin 
America’s sales tax or VAT systems is seen as a liability.  Argentina, however, 
provides the best-case scenario in this analysis because it has already started to 
switch over to a destination-based subnational VAT system (Table 3). 
 
Brazil’s problematic VAT-system specifically, can therefore provide some 
interesting insights for the future decentralisation exercises of individual SADC 
members or for further integration of the SADC members (subnational level) 
into a common market or an economic union.  Brazil was the first country to 
introduce a fully-fledged VAT in 1967.  The introduction of VAT in Brazil and 
the consequent problems are directly related to the fact that it was also 
introduced as a subnational VAT.  The rate structure also changed from a single 
to a multiple rate system and the new tax resulted in a series of complex 
technical and administrative problems of how to apply different VATs in the 
different states (the ICMS for each state) in addition to a federal VAT (the IPI)8.   



SAJEMS NS Vol 7 (2004) No 2 415

At present, the origin principle applies to interstate trade in Brazil.  There is no 
meaningful concept of administrative integration between the federal and state 
versions of the VAT.  Brazil therefore has the problems of dealing with cross-
border trade which has been problematic even in the EU, but also excessive 
compliance and administrative costs, location distortion, and tax exporting and 
competition.  Despite uniform rates for the states on exports and attempts to 
alleviate the distortionary effect of the origin principle by imposing a standard 
rate of 12 per cent on interstate trade (with an exception of a lower rate of 7 per 
cent on shipments to the poorer state), there are still economic complications. 
 
Recent recommendations in Brazil made provision for the adoption of an 
integrated VAT system with a new federal ICMS that would be collected 
together with a revised state ICMS on the same base as a unified VAT at a 
uniform national rate consisting of a federal rate and a uniform state rate similar 
to the “harmonised” VAT system in Canada or the proposed “common” VAT 
system (1996) in the European Commission.  After considerable debate, the 
state ICMS was substantially revised to eliminate significant elements of 
taxation on exports and investment in the existing system, with the federal 
government guaranteeing that no state would lose revenue as a result of the 
change.  It is therefore argued that in Brazil, as in the case of Argentina and 
India, a decent VAT system with subnational governments also applying 
independent VATs will require the implementation of the destination principle 
at different tax rates on interstate trade and some means of compensating 
“losing” states for revenue losses implied by the transition.  This type of system 
resembles the one in Canada, which is also similar to the CVAT option and also 
the tax-sharing option (specifically the so-called gewerbesteuer or local 
business tax) in Germany.   
 
A good administrative system assisted by mutual trust and a high degree of 
negotiation between the different levels of government is necessary when the 
destination principle is applied.  The theoretical case for the destination 
(residence principle) is strong but not absolute.  The ease with which some 
commodities or capital goods can be moved means that a significant element of 
origin (source) taxation is always inescapable.  However, consensus seems to 
exist in favour of maintaining as much of the destination principle as possible 
(maybe supporting it by use of restrictions on distance sales); also because of a 
fear of transfer pricing problems that potentially arise when VAT is levied by 
the origin principle9.       
 
Although SADC governments’ administrative capacities seem to have improved 
since the 1980s (Shalizi & Squire, 1988), a high degree of mutual trust such as 
the one present in the “harmonised” VAT system of Canada or tax-sharing 
options of Germany, could still elude these governments10.  The growing 



SAJEMS NS Vol 7 (2004) No 2 416 

problem of direct sales through electronic commerce might still also become a 
problem in the SADC.  The Green Paper on E-Commerce (RSA, 2000b: 30) in 
South Africa also recognises this problem and questions to what extent e-
consumption should be taxed.  A consideration of alternative approaches (the 
CVAT and VIVAT systems) to the destination-based VAT system has therefore 
become necessary11.   
 
When considering alternative approaches, the importance of an overarching 
authority again becomes clear.  Keen (2000) argues that in the absence of an 
overarching authority, VAT systems run into difficulty in securing appropriate 
clearing, ensuring that revenue collected on exports from one jurisdiction is 
available to finance credits/refunds claimed in another.  This problem could be 
resolved by providing incentives to subnational tax administrations to provide 
the appropriate level of effort in terms of their wider collective interests.  Of 
course, also in terms of the EU where tax sovereignty is regarded as more 
important, the adoption of an overarching federal or supra-national authority 
becomes a natural outflow. 
 
In the SADC the harmonisation or coordination of commodity taxation could 
become necessary in the long term.  In the short term it is important to 
concentrate on revenue losses and therefore rate increases for some countries, 
notably Zambia and Zimbabwe.  Cross-border trading has also shown 
significant growth with South African cross-border debtor finance worth 
approximately $20 million in April 200112.  With enhanced cross-border trade 
the possibility of factor movements increases and the concept of tax competition 
in terms of customs and excise duties, sales taxes and VAT becomes relevant.  
In this regard, South African authorities would have to take the future effect of 
the enhanced taxing powers for South African provinces into account in terms 
of the whole SADC region.  Intergovernmental relations will have to be 
coordinated in accordance with a strategy for the whole of the SADC.   
 
The adoption of the destination principle is therefore advisable for a future 
SADC, although not always administratively feasible, and could secure most of 
the neutrality needed in a region that is in a process of trade liberalisation.  With 
deeper integration and in the absence of border controls, a national accounts 
clearance mechanism operated by SARS or an agreed upon independent 
revenue authority, could be the sensible route to follow.  Even if the destination 
principle prevails in the SADC, governments must consider the adoption of 
permissible tax rate “bands” for VATs although the benefits of complete 
harmonisation are unlikely to exceed the costs (CREFSA, 1998).  A degree of 
flexibility could drive a “healthy”, export-oriented competitive process with 
automatic harmonisation.  
 



SAJEMS NS Vol 7 (2004) No 2 417

Faria (1995: 24) summarises the experience of the EU with VAT as follows:  
“Within the EU, it has proved easier, in relation to VAT, to agree on the nature 
of the tax (consumption or destination type), and the base (virtually all domestic 
consumption goods and services except investment goods or financial services) 
than the tax rate structure (number and levels of rates, although the 6th Directive 
has formalised a minimum rate level of 15 per cent)”.  Tax sovereignty 
(normally secured by the origin principle) will probably also be high on the 
agenda of SADC countries.  Although a spontaneous harmonisation is possible 
with the adoption of proper convergence criteria as in the case of the EU, an 
overarching (independent) fiscal authority (and maybe a clearing mechanism in 
the absence of border controls) may be necessary for administrative ease and the 
perfection of VAT operation.  This can only be achieved through a high degree 
of close cooperation and negotiation on the SADC governments’ part, especially 
through Ministers of Finance.  In short, developing countries such as those in 
the SADC have to choose a tax system that promotes growth and development.   
 
Another factor that should never be overlooked in the choice of the most 
appropriate system of taxation in a future SADC is that a balance should be 
maintained between commodity and capital income taxation.  It could happen 
that one country is in favour of high commodity taxation but not capital income 
taxation.  The opposite could be true of another country within a future SADC, 
with the argument that low commodity taxation compensates for high capital 
income taxation.  In the first case, the country with the high commodity taxation 
would favour the destination principle (with border tax adjustments).  In the 
second, the country with the low commodity taxation would favour the origin 
principle because exports already carry a high capital income (corporate) tax 
burden.   
 
 
5 CONCLUSION AND RECOMMENDATIONS 
 
Most SADC members have centralised or unitary systems with some 
characterised by authoritarian rule and/or high military expenditures especially 
in war-torn countries such as Angola and the DRC.  Several heads of state such 
as those of Namibia and Uganda, with the most recent case in Zambia, have also 
opted to adjust their constitutions in order to lengthen their terms of office.  
Data are therefore not always available or nonexistent for subnational levels.  In 
a number of countries, decentralisation has not yet resulted in a relinquished 
control from the centre, and is partly related to the quality of governance at 
different levels.  Besides Uganda, South Africa is regarded as one of the few 
African countries that are in a process of unification through decentralisation.  
The role of a democratic South Africa in terms of further regional integration 
(and thus the process of fiscal decentralisation) in Southern Africa therefore has 



SAJEMS NS Vol 7 (2004) No 2 418 

to be confirmed.      
 
Most SADC members will have to take action to restore internal balance, 
although countries such as Botswana, Lesotho, Mauritius, Namibia, Swaziland 
and Seychelles are already showing signs that further trade liberalisation will be 
to their advantage.  Although countries such as Malawi, Mozambique, Zambia 
and Zimbabwe could lose in terms of revenue, broadened tax bases, improved 
tax compliance and rate increases of especially commodity taxes could 
significantly alleviate the problem.  The importance of strong leadership should 
also be recognised, and in this regard it is advisable for the SADC to utilise 
South Africa’s experience and resources to its advantage instead.   
 
The South African economy is, an exception especially because of its relative 
size in the SADC. However, this should be regarded as an advantage in the 
region.  The South African economy already makes significant contributions in 
terms of exports and imports in the region.  It can therefore be expected that 
although compensation may be needed in the shorter term, a range of benefits 
from an expansion of foreign trade over the longer term would make the SADC 
economies less dependent on South Africa.  A continuous process of trade 
liberalisation exercised with care could therefore be beneficial, although 
strategies such as tax competition (including profit-shifting) would have to be 
taken into account especially competition from other regions dominating trade 
in the developing world, for instance, Asia.  This also includes effective 
commodity tax competition and/or coordination in a future SADC.  Future 
studies could, therefore, model the behaviour of SADC countries in order to 
measure the degree of competition in terms of commodity taxation.  
   
Issues emphasised and recommended include the following:  (1) although not 
always administratively feasible, the destination principle can secure most of the 
neutrality needed in a region which is in a process of trade liberalisation; (2) 
directly linked to (1), the adoption of a national accounts clearance mechanism 
operated by SARS or an agreed upon independent revenue authority will be 
advisable with further integration; (3) governments could consider the adoption 
of permissible tax rate “bands” for VATs or minimum rates; although (d) a 
degree of flexibility (applied with care) could drive a “healthy” export-oriented 
competitive process with automatic harmonisation. 
 
 
 
ENDNOTES 
 
1 Argentina’s foreign debt soared to $149,7 billion in 2001. 



SAJEMS NS Vol 7 (2004) No 2 419

2 This tax is levied at producer’s level that is, a uniform tax is collected 
only on the output of domestic firms regardless of where this output is 
ultimately consumed.  Exports are taxable whilst imports are zero-rated 
(production tax).  Zero-rating implies that the firm files a return but pays 
zero tax on sales and gets a refund in respect of VAT payments made at 
earlier stages in the production and distribution chain.  By contrast, 
exemption implies the firm need not file a VAT return and does not levy 
VAT but also cannot claim refunds for any VAT included. 

3 The commodity tax is levied at the consumer’s level and enables the 
region to collect a tax on all of its residents’ private good consumption. If 
the destination principle is adopted in both the home and host regions, 
imports are taxed at the same rate as domestically produced goods and 
exports are zero-rated.  As in the case of the residence principle, the 
destination principle is perceived to be a more fair and equitable practice 
because domestic and imported goods are treated the same (exports are 
zero-rated but imports are taxable). Administrative problems may be 
more likely with this system because of information difficulties in the 
taxation of imported goods.     

4 Questions were also raised about the legitimacy of the elections in 2001 
in which the current president, Mr Museveni, was re-elected with nearly 
70 per cent of the votes.           

5 In the event of directing various requests for information from the 
different SADC members, it was interesting to find that these countries 
do not have information about one another’s tax systems and tax rates 
which indicates a lack of either cooperation or communication.  

6 Revenue productivity of commodity taxes such as VAT/GST can be 
measured as follows:  VAT/GST as  per cent of GDP: VAT/GST tax rate. 

7 In an electronic discussion with Mr GT Pasi, the Commissioner of Taxes 
from Zimbabwe, it was confirmed that Zimbabwe is also planning to 
switch over to a destination-based VAT.   

8 Nowadays, Brazil levies a VAT payable on sales and transfers of goods 
(industry) in the form of a federal excise tax or then VAT (IPI or Imposto 
Sobre Productos Industrializados) at various rates in accordance with the 
nature of the product (10 to 15 per cent and in certain cases over 300 per 
cent).  A state sales and services tax or VAT on agriculture, industry and 
other services (ICMS or Imposto Sobre Operacoes Relativas a 
Circulacao de Mercadirias e Servicios) of 7 to 25 per cent is also levied. 
In addition, a municipal services tax, the ISS (Imposto Sobre Servicios) is 
levied on gross income by municipalities on a variety of industrial, 
commercial, and professional services levied on gross income by 
municipalities on a variety of industrial, commercial and professional 
services.    

9 Levying VAT on an origin basis effectively means charging the value 



SAJEMS NS Vol 7 (2004) No 2 420 

that is added to a product in different jurisdictions at the rates charged by 
those jurisdictions.  Multinational firms or firms operating in multiple 
jurisdictions then have an incentive to transfer price value-added into 
low-tax jurisdictions, say, charging high internal prices for intrafirm sales 
out of them. 

10 In an electronic questionnaire to some of the SADC governments, one of 
the requests was whether they had any information about their 
neighbouring countries’ tax systems.  The answer was conclusively 
negative. 

11 See Keen (2000) for an extensive discussion of alternative approaches to 
the destination-based VAT system such as the CVAT proposal for India 
and Brazil where an over-arching federal government exists and the 
viable integrated VAT or VIVAT proposal for the EU where no such 
authority exists.  

12 International cross-border debtor finance was worth $500 billion in 2000. 
 
 
Acknowledgements:  The author would like to thank the anonymous referees for 
useful comments. 
 
 



SAJEMS NS Vol 7 (2004) No 2 421

APPENDIX A STATISTICAL TABLES 
 
Table A.1 Sub-national shares of selected federations and the EU  
 

Share of total government 
expenditure (%) 

 
Share of tax revenue (%) 

 
Country 

1990 1997-2001 1990 1997-2001 
Australia1 50,4 49,7 20,7 25,1 
Austria1 31,9 32,2 21,7 20,7 
Belgium1 11,9 11,8 4,5 5,4 
Canada1 58,7 49,4 49,5 43,5 
Denmark 54,8 54,5 31,1 31,5 
Finland 46,5 41,2 25,9 27,6 
France 18,7 18,6 9,7 10,8 
Germany1 40,2 37,8 28,9 28,8 
Greece --- 7,9 (1994) --- 3,6 (1994) 
Ireland 27,9 30,7 2,5 2,4 
Italy 22,8 25,4 3,6 6,5 
Luxembourg 19,9 16,9 6,8 6,3 
The 
Netherlands 

29,0 26,1 3,4 4,1 

Portugal 8,7 11,6 3,6 5,9 
Spain 34,3 35,0 13,3 13,8 
Sweden 39,8 36,2 28,2 31,4 
Switzerland 51,2 49,3 37,0 35,5 
UK 29,0 27,0 5,9 3,6 
US1 42,0 46,4 33,8 32,9 

Note: These countries represent federations whilst the rest are EU members 
(including Austria, Belgium & Germany)    

Source:  Own calculations - IMF (2002)   
 



SAJEMS NS Vol 7 (2004) No 2 422 

Table A.2 Commodity tax revenue, 2002 
 

Commodity taxes 
as % of GDP 

Commodity taxes as 
% of total taxation Country 

A1 B1 A B 

Total 
taxation 
as % of 

GDP 
Canada (provinces) 2,7 1,6 21,2 12,5 12,5 
US (states) 1,9 0,9 33,3 16,0 5,6 
Austria2 6,0 2,7 27,2 12,2 33,4 
Belgium 1,3 3,2 8,1 19,8 33,2 
Denmark 10,2 5,8 29,8 17,0 35,3 
Finland 8,6 5,4 35,4 22,2 36,0 
France 7,9 3,3 37,9 16,1 40,1 
Germany 3,4 3,2 30,1 28,0 27,9 
Greece 9,4 6,6 33,7 23,5 40,3 
Ireland 7,0 5,2 24,8 18,3 33,0 
Italy 5,6 4,1 21,1 15,4 39,3 
Luxembourg 6,7 5,3 22,4 17,8 41,3 
The Netherlands 7,0 3,5 28,7 14,5 43,4 
Portugal 7,1 6,2 30,4 26,7 34,5 
Spain 4,9 2,9 28,7 16,7 41,3 
Sweden 7,0 4,4 29,7 18,8 33,4 
UK 6,4 4,7 23,2 17,2 34,2 
EU 15 average3 6,6 4,4 27,4 18,9 36,4 

Notes:  1 Commodity taxation includes categories A and B, that is general     
taxes (including VAT) and taxes on specific goods and services 
(including excise taxes) respectively as defined by the OECD.    
2 Starting with Austria, the EU members’ commodity taxation is shown 
at central or national level. 
3  Unweighted. 

Source:  OECD (2003) 
 
 
 
 
 



SAJEMS NS Vol 7 (2004) No 2 423

REFERENCES 
 
1 BIRD, R.M. (1992) “Tax reform in Latin America: A review of some 

recent experiences”,  Latin American Review, 27: 7-36. 
2 CNOSSEN, S. (1998) “Reforming company taxation in Europe”, in 

Sorenson, P.B. (ed.) Public Finance in a Changing World, MacMillan: 
London: 221-54. 

3 COHEN, T. & STEYN, G. (2001) “De Beers needs special meeting”,   
Business Day 24 April. 

4 CREFSA, see Centre for Research into Economics and Finance in 
Southern Africa. 

5 DE BONIS, V. (1997) “Regional Integration and commodity tax 
harmonization”, World Bank Policy Research working paper 1848.    

6 DE CROMBRUGGHE, A & TULKENS, H. (1990) “On Pareto 
improving commodity tax changes under fiscal competition”,   Journal of 
Public Economics, 41: 335-50. 

7 DEPARTMENT OF TRADE AND INDUSTRY (2001) South African 
Trade Statistics   (http://www.dti.gov.za). 

8 EC, see European Communities Commission. 
9 EUROPEAN COMMUNITIES COMMISSION (1987) Completing the 

internal market: The introduction of a VAT clearing mechanism for intra-
Community sales,  EC: Brussels. 

10 ________ (1988) The Economics of 1992: An Assessment of the Potential 
Economic Effect of Completing the Internal Market in the European 
Community,  EC: Brussels.  

11 ________ (1996)  “A common system of VAT:  A programme for the 
single market”, COM 96/328 final. 

12 EVANS, H.D. (1997) “Study of the impact of the removal of tariffs of the 
free trade area of the Southern African development community 
(SADC)”, Occasional Paper No. 5, Commonwealth Secretariat, Pretoria: 
SADC Industry and Trade Coordination Division (SITCD). 

13 ________ (1998) The regional trade model of Southern Africa: New data 
and results,   Final Report,   DTI and IDC: Pretoria. 

14 ________ (2000) Options for regional integration in Southern Africa,   
South African Journal of Economics, 68(4): 662-92. 

15 FARIA, A.G.A. (1995) “Tax coordination and harmonization”, in Shome, 
P (ed.) Tax Policy Handbook, IMF: Washington D.C. 

16 FEENBERG, D.R., MITRUSI, A.W. & POTERBA, J.M. (1997) 
“Distributional effects of adopting a national retail sales tax”, in Poterba, 
JM (ed.) Tax Policy and the Economy, Vol 11, MIT Press: Cambridge 
MA: 49-90. 

17 FEHR, H., ROSENBERG, C. & WIEGARD, W. (1995) Welfare Effects 
of Value Added Tax Harmonization in Europe,  Springer: Berlin. 



SAJEMS NS Vol 7 (2004) No 2 424 

18 FINANCIAL AND FISCAL COMMISSION, see South Africa 
(Republic). 

19 FINANCE AND INVESTMENT SECTOR CO-ORDINATING UNIT 
(1999)  (www.fiscu.gov.za). 

20 FITZGERALD, J., QUINN, T., WHELAN, B. & WILLIAMS, J. (1988)   
“An analysis of cross-border shopping”, Paper 137, The Economic and 
Social Research Institute: Dublin. 

21 HAUFLER, A. (1993) Commodity Tax Harmonization in the European 
Community: A General Equilibrium Analysis of Tax Policy Options in the 
Internal Market,  Physica-Verlag: Germany. 

22 ________ (1998) “Asymmetric commodity tax competition: Comment on 
De Crombrugghe and Tulkens”, Journal of Public Economics 67: 135-44. 

23 IMF, see International Monetary Fund. 
24 INTERNATIONAL MONETARY FUND (2000) “Argentina: Selected 

issues and statistical annex”, IMF Staff Country Report No. 00/160, IMF: 
Washington D.C. 

25 ________ (2001) “Canada:  Selected Issues”, IMF Staff Country Report 
No. 01/157,  IMF: Washington D.C. 

26 ________ (2002) Government Finance Statistics Yearbook, IMF: 
Washington D.C. 

27 KATZ COMMISSION (1994) Interim Report of the Commission of 
Inquiry Into Certain Aspects of the Tax Structure of South Africa,  
Government Printer: Pretoria. 

28 ________ (1995) Third Interim Report of the Commission of Inquiry Into 
Certain Aspects of the Tax Structure of South Africa, Government Printer: 
Pretoria. 

29 KEEN, M. (1987) “Welfare effects of commodity tax harmonization”,  
Journal of Public Economics, 33: 107-14. 

30 ________ (1989) “Pareto-improving indirect tax harmonization”   
European Economic Review 33: 1-12. 

31 ________ (2000) “VIVAT, CVAT and all that: New forms of value-
added tax for federal systems”, IMF Fiscal Affairs Department: 
Washington D.C.: 1-13.  

32 KEEN, M. & LIGTHART, J.E. (1999) “Coordinating tariff reduction and 
domestic tax reform”, IMF Working Paper WP/99/93,  IMF: Washington 
D.C. 

33 LEVINSOHN, J. & SLEMROD, J. (1993) “Taxes, tariffs and the global 
corporation”,  Journal of Public Economics, 51: 97-116. 

34 LOCKWOOD, B. (1993) “Commodity tax competition under destination 
and origin principles”, Journal of Public Economics, 52: 141-62. 

35 MARGO COMMISSION (1987) Report of the commission of inquiry 
into the tax structure of the Republic of South Africa,  Pretoria: South 
Africa. 



SAJEMS NS Vol 7 (2004) No 2 425

36 McCCARTHY, C.L. (1999) “Polarised development in a SADC free 
trade area”, South African Journal of Economics, 67(4): 375-99. 

37 THE MILLENIUM PARTNERSHIP FOR THE AFRICAN RECOVERY 
PROGRAMME (MAP) (2001)  Government Printer: Pretoria. 

38 MURPHY, R.L. & MUSCOVITS, C. (1999) “Decentralisation, inter-
governmental fiscal relations and macroeconomic governance:  The Case 
of Argentina”, in Fukusaku, K & de Mello, LR, Fiscal Decentralisation 
in Emerging Economies, Development Centre (OECD): Paris: 121-34. 

39 NAUDE, W.A. & KRUGELL, W. (2001) “Global integration of Africa 
versus regional integration in Africa”, South African Journal of Economic 
and Management Sciences 4(3): 493-504. 

40 NATIONAL TREASURY OF SOUTH AFRICA. (2000) The budget 
review, NTSA: Pretoria. 

41 ________ (2001)   The budget review,   Pretoria: NTSA. 
42 ________ (2003)   The budget review,   Pretoria: NTSA. 
43 NTSA, see National Treasury of South Africa. 
44 OMEGA PLAN (2001) Dakar:  Senegal. 
45 ORGANISATION FOR ECONOMIC COOPERATION AND 

DEVELOPMENT (1994) Tax information exchange between OECD 
member countries: A survey of current practices, OECD Publications: 
Paris. 

46 ________ (1998) Harmful tax competition: An emerging global issue,   
Paris: OECD Publications. 

47 ________ (2003) OECD Revenue Statistics,   Paris: OECD Publications. 
48 PRICEWATERHOUSECOOPERS (PwC) (2004) Corporate Taxes 2003-

2004.  Worldwide Summaries, New York: John Wiley and Sons. 
49 RSA, see SOUTH AFRICA (Republic) 
50 SADC, see Southern African Development Community. 
51 SARB, see South African Reserve Bank. 
52 SHAH, A. (1999) “Fiscal federalism and macroeconomic governance: 

For better or for worse?” in Fukasaku, K. & Mello, L.R. (eds.) Fiscal 
Decentralisation in Emerging Economies: Governance Issues, OECD 
Publications: Paris. 

53 SHALIZI, Z. & SQUIRE, L. (1988) “Tax policy in Sub-Saharan Africa”,   
Policy and Research Series No. 2, World Bank Publications: Washington 
D.C. 

54 SOUTHERN AFRICAN DEVELOPMENT COMMUNITY  2001-2004.  
SADC member inputs, Telephonic and electronic interviews, April:  
Pretoria. 

55 SOUTH AFRICA (Republic) (1996) The Financial and Fiscal 
Commission’s Recommendations for the Allocation of Financial 
Resources to the National and Provincial Governments for the 1997/98 
Financial Year (http//www.budget.gov.za). 



SAJEMS NS Vol 7 (2004) No 2 426 

56 SOUTH AFRICA (Republic) (1997) The constitution of the Republic of 
South Africa,   Wynberg: Hn Communications.  

57 ________ (Republic) (2000) Green Paper on E-Commerce:  Making it 
your Business.   Pretoria: Department of Communications.       

58 SOUTH AFRICAN RESERVE BANK  Quarterly Bulletin of the Reserve 
Bank,   Various Issues. 

59 STEIN, E. (1998) “Fiscal decentralisation and government size in Latin 
America”, in Fukasaku, K. & Hausmann, R. (eds.) Democracy, 
Decentralisation and Deficits in Latin America, OECD Publications: 
Paris: 95-120.   

60 TAIT, A.A. (1988)  Value Added Tax,  IMF: Washington D.C. 
61 WORLD ECONOMIC FORUM (WEF) (2003) The Global 

Competitiveness Report: Switzerland.   
62 ZEE, H.H. (1995) “Value added tax”, in Shome, P (ed.) Tax Policy 

Handbook, IMF: Washington D.C.