The extent of anti-export bias in the South African economy during the 1990s

Abstract

An analysis of trade incentives in the South African economy during the 1990s reveals two major findings. Firstly, the extent of the anti-export bias in South Africa's trade policy during the 1990s is less than is claimed in the empirical literature. The sectors subjected to an anti-export bias accounted for around 7 per cent of total output in 1990 and 21 per cent in 1999. Secondly, sectoral output growth did not strongly correlate with the prevailing trade incentives of the 1990s. Export production continued despite the prevalence of import substituting incentives in many sectors. Although further research is needed, this suggests that attention should also be given to other factors related to domestic competitiveness (e.g. skills development, productivity enhancement, competition policy, etc.) and market access if South African export production is to be significantly increased.

1. Introduction

It has been shown that in a three-sector framework the promotion of import substituting production need not be at the expense of export production, and vice versa (Sachs, 1985; Pack & Westphal, 1986; Singer & Alizadeh, 1986; Liang, 1992). This is an important development since it calls into question the conventional measure of anti-export bias, namely that import substitution can only be at the expense of export production. Hence the first objective of this paper is to measure the extent of anti-export bias in the South African economy during the 1990s within a three-sector framework and the second is to ascertain whether the actual trade patterns conform to the policy incentives accorded to industries during this period.

Section 2 provides a theoretical review of trade incentives and their impact on trade regime bias, Section 3 an empirical analysis of the bias of the trade regime in South Africa during the 1990s, Section 4 a brief review of manufacturing production in the light of the trade regime bias identified in the previous section, and Section 5 concludes.

2. Trade incentives

The general case for trade strategy as the main determinant of industrial success is based on the assumption that incentives are an important determinant of performance (Lall, 1990: 119). However, it is important to recognise that while trade policy (for example trade incentives such as tariffs, quotas, export subsidies, etc.) is important, other important elements of industrial policy include tax policy, employment policies, competition policies, research and development (R&D), policies influencing technology transfer and growth of domestic markets. The process of industrialisation is to a large extent determined by the interplay between these different elements. Trade incentives that encourage export production, for example, may not be successful if they are not complemented by policies that ensure favourable access to credit, contribute to human resource development, promote R&D activity, etc. For example, technology-focused theories predict that the magnitude, direction and gains from trade arise from technology gaps between countries (Posner, 1961; Hufbauer, 1966; Vernon, 1966; Krugman, 1979). The point being emphasised here is that since industrial policy is a multidimensional process one should be careful not to downplay or overstate the importance of trade policy in the industrialisation process without due consideration of all the other elements/policies affecting industrialisation – or at least the important ones. However, this concern is not of direct relevance here since this paper is primarily concerned with ascertaining the extent of anti-export bias in the South African economy during the 1990s.

Measuring trade incentives has been one of the major challenges confronting the empirical analyst. This challenge stems from the non-availability of reliable data owing either to deficiencies in statistical records or information being deliberately excluded from official statistics. Information could be excluded from official records in order to keep domestic lobby pressures in check, to avoid falling foul of World Trade Organization (WTO) rules, or for strategic purposes. This last point is of particular relevance for South Africa – during the apartheid era trade in some products (most notably gold and oil) was regarded as being highly strategic and hence statistics were not always released.

Trade incentives could include direct measures such as tariffs, quotas and export subsidies and indirect measures such as special tax incentives to promote production (as in the case of export processing zones) and expenditures on R&D and skills development. Information on import restrictions (e.g. tariffs) and import quotas is usually more readily available and these have mainly been used in the appraisal of trade policy and incorporated into effective protection analysis (for the application to South Africa see Holden & Holden, 1978; IDC, 1996a; Fedderke & Vaze, 2001). However, if export and import substituting incentives were simultaneously used to stimulate production, then a critical analysis of trade policy has to analyse both sets of incentives. Between 1990 and 1997, both import protection and export subsidies (under the General Export Incentive Scheme) were given to sectors – the issue of relevance is what effect these incentives had on the trade regime bias during the 1990s.

Even if a realistic measure of trade incentives can be derived it is still necessary to define the criteria determining the trade policy stance. In other words, what is the level of trade incentives that biases the regime towards either export or import competing production? Krueger (1978) defines the overall stance or ‘bias’ of trade policy as the ratio of the internal relative price of exports and imports (internal terms of trade) to the world price ratio (external terms of trade). Expressed mathematically this is given by:

where TB reflects the trade bias, P _x and P_m the domestic price of exports and imports and and world prices of exports and imports. TB will exceed unity when the domestic incentive structure promotes exports. A value less than unity depicts an anti-export bias since import substituting production is being favoured. The incentive structure is neutral when TB equals 1. Equation 1 provides an indirect measure of the trade incentives accorded to export or import competing production. However, a major problem associated with equation 1 as a measure of trade incentives relates to the availability of accurate measures for both domestic and world prices.

Bhagwati (1988) defines export promotion as a strategy that does not discriminate against exports. By this definition, a neutral trade policy stance would qualify as export promotion, and an incentive structure that favours imports over exports would be construed as having an anti-export bias. Bhagwati uses the effective exchange rate (EER) to depict the bias in the trade regime. In this case, the EER is made up of the nominal exchange rate plus any trade incentives per unit of foreign currency received by domestic producers. This can be represented as follows:

where e, t _m , q _m , s _x represent the exchange rate, the tariff rate, the ad valorem equivalent of the quota and the export subsidy rate respectively. In addition, t _x represents any disincentive (e.g. an export tax) against exports. The trade regime is said to be neutral if EER _x equals EER _m . The EER can also be calculated at a sectoral level, in which case equation 1 reflects the bias of the trade regime at the sectoral level.

In the 1960s and 1970s the trade policy debate on economic development centred on protection for import substitution vis-à-vis export promoting activities, with the Latin American experience providing fertile ground for this debate. Since the mid-1980s it has become apparent that the main issue involved in the path to industrialisation has not been one of either export promotion or import substitution but the inter-relationship between these two strategies and their sequencing. The notion that protection restricts export growth is based on the two-sector (exportables and importables) model where protection of one sector is at the expense of the other sector (Clements & Sjaastad, 1984; Greenaway & Milner, 1987; World Bank, 1987). A conventional measure of the anti-export bias is given by:

where AEB = anti-export bias

VA _m  = value added of import products under protection,  = value added of import products under free trade, VA _x  = value added of export products under protection,  = value added of export products under free trade

According to equation 4, if AEB exceeds 1 then the incentive to produce for the domestic market outweighs the production for the export market – stated differently, there is a bias against export production. Since equation 4 is based on a two-sector framework, any import substituting incentives is at the expense of export production, thus implying an anti-export bias in the trade regime. This can be illustrated quite easily as follows. If we assume that there are no incentives for export production, but incentives for domestic production result in equalling 0.1 (10 per cent) and VA _m equalling 0.2 (20 per cent), then this results in an AEB measure of 2.

However, in a three-sector model, even with full employment, import substituting policies can complement export promoting policies as resources are drawn from non-tradables into both the tradables sectors (Sachs, 1985; Pack & Westphal, 1986; Singer & Alizadeh, 1986). Liang (1992) has shown that in a three-sector model (exportables, importables and non-tradables), export promotion and import substitution need not be mutually exclusive policies. In this case production of tradables is a function of two relative prices, that is:

An increase/decrease in the price of exportables (P _x ) results in an increase/decrease in the production of exportables. However, the increase/decrease is not necessarily at the expense/to the advantage of the import substituting sector. This is because the increase in the production of tradables (exportables, importables) can be facilitated through a shift of resources from the non-tradables to the tradables sector. Using this three-sector framework it is possible to identify five distinct trade incentive patterns as reflected in Figure 1 (Liang, 1992).

• Quadrant 1 reflects a ‘pure’ export promoting strategy (EP) where export incentives are positive and negative protection (disincentives) for import substituting activities.

• In quadrant 2 there are incentives for both export activities and import substituting activities. Liang (1992) terms this ‘protective export promotion’ (PEP). In this case, protection is accorded for the domestic market while firms are simultaneously encouraged to export. This corresponds with South Korea's early export experience where incentives for both export and import substituting activities were simultaneously provided (Suh, 1975; Pack & Westphal, 1986).

• Quadrant 3 depicts a situation where disincentives exist for both exportables and import substitutes. Imports and non-tradables are being favoured.

• Quadrant 4 depicts ‘import substitution’ (IS) where there are incentives for import substitutes and disincentives for exportables.

• A neutral trade policy stance is one where neither exportables nor import substitutes receive any incentives. This is captured at the point of intersection of the two axes.

Thus, within a three-sector framework, free trade is but one of a range of export-fostering policy regimes. The simultaneous protection of both the exportables and importables sector is not incompatible with export promotion. Wade (1990, 1992) and Amsden (1989) have argued that the East Asian experience has shown that import protection was necessary to secure export production. In addition, Krugman (1994) has shown that under conditions of imperfect competition import protection is not only compatible with but may also be necessary for export production. The implication of this is that it calls into question the conventional interpretation that a greater incentive to produce for the domestic market is a bias against export production. In terms of Figure 1, only quadrants 3 and 4 reflect a bias against exports. Quadrant 2, although having incentives for import substitution, also incorporates incentives for export production and hence does not depict a bias against exports. This is the fundamental point that emerges from an analysis within a three-sector framework: import substitution need not be at the expense of export promotion.

Figure 1: Classification of trade incentives

However, Milner (1995) has shown that the simultaneous promotion of exportables and importables may not necessarily produce a pro-tradable bias. The net effects depend on the nature and magnitude of the substitution, complementarity and income effects of the exportables, importables and non-tradable sectors (Milner, 1995). In addition, the existence of imperfect competition may result in policy measures not matching production outcomes. Since actual outcomes may differ from policy intentions, Liang's (1992) trade incentive classifications may have limitations as an ex ante tool of policy formulation. However, his classifications provide a useful tool for ex post analysis, allowing one to at least ascertain whether the bias of the trade regime matched policy intentions.

Following the general equilibrium framework developed by Sjaastad (1980) and Greenaway and Milner (1987), incidence analysis has also been used to analyse the effects of protection. In a three-sector model:

an examination of how an import tariff alters the price of importables relative to exportables and non-tradables can provide an indication of the ‘true’ protection of importables and the extent to which the incidence of the tax is shifted onto exportables and non-tradables. The incidence depends essentially on the degree of substitutability (in demand and production) between the products of the importables sector and the other unprotected sectors. (Greenaway, 1989: 127)

The incidence measure is depicted by the variable ‘w’ in the following formula (Greenaway, 1989; Milner, 1995):

where u is the stochastic disturbance term. In this case w estimates the proportion of import protection that is shifted in the form of an implicit export tax. Where importables and non-tradables are substitutes, w tends towards unity and w tends towards zero if exportables and non-tradables are substitutes for each other (Greenaway, 1989).

Another model, which unfortunately has strong data requirements, is the trade restrictive index (TRI) proposed by Anderson and Neary (1996). The (TRI) uses a computable general equilibrium (CGE) model to derive the uniform tariff, which has the same static welfare effect as the structure of tariffs and quotas actually in place. Similarly, a trade subsidisation index can also be constructed to capture the effects of export subsidies (Anderson & Neary, 1996). The data requirements for the calculation of the trade restrictive index preclude its use in the case of South Africa. In addition, O'Rourke (2000) has shown that the index is sensitive to changes in both the model specification and demand elasticities.

3. Trade regime bias in South Africa

As in the case of Balassa and Associates (1982), we use the effective rate of protection to reflect the trade incentives accorded to import substituting activities. The ERP rates were sourced from Fedderke and Vaze (2001). This is termed the net effective subsidy rate on imports. The net effective subsidy rate on exports captures the combined effects of protective measures and export incentives on export production. In the calculation of export incentives we took account of the export subsidies under the General Export Incentive Scheme (GEIS), import rebates and the tariffs paid on intermediate inputs. The import rebates included those available under Article 470.03 of the Customs and Excise Act. Owing to data constraints, transport rebates and interest rate concessions that may have been accorded to some industries during the period under analysis were not considered. Balassa and Associates (1982: Appendix 1) provide a detailed description of how the trade incentive bias can be measured. The sectors considered in the analysis of trade regime bias are reflected in Table 1.

Table 1:  Sectors reflected in Figures 2 and 3

1	Agriculture, forestry and fishing	17	Other chemical & man-made fibres
2	Coal mining	18	Rubber
3	Gold and uranium ore mining	19	Plastic products
4	Other mining	20	Glass and glass products
5	Food	21	Non-metallic minerals
6	Beverages	22	Basic iron and steel
7	Tobacco	23	Basic non-ferrous metals
8	Textiles	24	Metal products excl. machinery
9	Wearing apparel	25	Machinery & equipment
10	Leather	26	Electrical machinery
11	Footwear	27	TV radio and equipment
12	Wood and wood products	28	Professional and scientific equipment
13	Paper and paper products	29	Motor vehicles parts
14	Printing, publishing and recording	30	Other transport
15	Coke and refined petroleum	31	Furniture
16	Basic chemicals	32	Other industries

Figures 2 and 3 present the trade regime bias of the various sectors for the periods 1990–94 and 1995–97. Since the GEIS began in 1990 and ended in 1997, and 1995 represented the beginning of SA's tariff liberalisation programme, the analysis is divided between the two periods 1990–94 and 1995–97. As in Liang (1992), a cut-off point of 5 per cent is used to define the free trade region. One should recognise that this is a subjective benchmark and as such influences the number of sectors that are classified as having a free trade regime bias. If a sector's incentive measure is within five percentage points of the intersection of the axis then the sector is defined as following a free trade strategy. The sectors are reflected in Table 1. The services sectors are excluded from the analysis on the assumption that they are mainly non-tradable sectors. Only those sectors that were not subjected to a free trade strategy are explicitly reflected in Figures 2 and 3. The sectors are represented by numbers (see Table 1) in Figures 2 and 3. All those sectors that are not reflected in Figures 2 and 3 were subjected to a free-trade regime bias.

Figure 2: Trade regime bias (1990–94). Source: Author's calculations with data from the Department of Trade and Industry

Figure 3: Trade regime bias (1965–97). Source: Author's calculations with data from the Department of Trade and Industry

During both periods, there were 17 sectors that fell outside the free trade region. For the period 1990–94, 12 sectors (metal products; food; furniture; TV, radio and communication; printing and publishing; plastic products; rubber products; leather and leather products; paper and paper products; basic iron and steel; footwear; glass and glass products) enjoyed both export promotion and import substituting incentives. As mentioned earlier, in terms of the classification used by Liang (1992) the incentive structure was one of ‘protected export promotion’ (PEP). This is not surprising given that both export subsidies and tariff protection were the two main trade policy instruments during this period. Four sectors (textiles; basic non-ferrous metals; professional and scientific equipment; tobacco) were given incentives/disincentives for import substitution/export production; this is classified as an import substituting trade strategy. There were disincentives for both export production and import substituting activities for the ‘other mining’ sector – this is classified as ‘de facto import promotion’ (DIP) – which is not surprising since ‘other mining’ mainly includes oil imports. Thus, during the period 1990–94 only five sectors (textiles; basic non-ferrous metals; professional scientific equipment; tobacco and other mining) had a trade policy bias against exports.

During the period 1995–97, export incentives (GEIS) were being phased out, with the result that only five sectors (professional and scientific equipment; food; basic iron and steel; rubber products; leather and leather products) enjoyed PEP. The number of sectors with an import substituting trade policy bias (i.e. incentives for import substitution and disincentives for export production) increased to 11. This included the following industries: electrical machinery and apparatus; motor vehicles and parts; agriculture; wearing apparel; paper and paper products; plastic products; printing and publishing; textiles; tobacco; footwear; glass and glass products. Imports were still being encouraged for the ‘other mining’ sector during this period.

Comparing the two periods, the following emerges:

• Taking into account all the sectors that had some bias in their trade regime during the periods under consideration, the incentive scheme is unchanged for four PEP sectors (food; leather and leather products; rubber products; basic iron and steel), one DIP sector (other mining) and two import substituting sectors (tobacco; textiles). Of the remaining sectors, four (footwear; printing and publishing; plastic products; glass and glass products) moved from enjoying both export and import substituting incentives during 1990–94 (PEP) to a situation where they were accorded incentives only for import substitution and disincentives for export production (‘import substitution’). The professional and scientific equipment sector moved from a situation of only having incentives/disincentives for import substitution/export promotion during 1990–94 to enjoying incentives for both export and import substituting activities during 1995–97. Four sectors (basic non-ferrous metals; metal products excluding machinery; TV, radio and communication; furniture) moved from enjoying some level of incentives for import substitution (e.g. basic non-ferrous metals) and some level of incentives for both export and import substituting activities (e.g. metal products excluding machinery; TV, radio and communication; furniture) during the period 1990–94 to a free trade regime bias during 1995–97. In addition, four sectors (agriculture; wearing apparel; electrical machinery; motor vehicles and parts) moved from a free trade regime bias during 1990–94 to an import substituting bias during 1995–97.

• There was an increase in the number of sectors subjected to an import substituting trade regime bias from four (1990–94) to eleven (1995–97). In essence, it is these eleven sectors that had an anti-export trade policy bias during the period 1995–97.

The significance of the last point should not be underestimated, or for that matter overestimated. It has been claimed that South Africa's trade policy has been characterised by a high level of anti-export bias (IDC, 1996a; Tsikata, 1999). This claim has been based on calculations undertaken within a two-sector framework – for example the IDC (1996a), identifies a large number of sectors that were subjected to an anti-export trade policy bias (see Appendix 1). In terms of these classifications, sectors subjected to an anti-export bias accounted for 87 per cent of total output in 1990 and 96 per cent in 1999 (see Appendix 3). The anti-export bias calculations by the IDC (1996a) for 1993 were used to determine the output contribution for the year 1990. The share of manufacturing sales subjected to an anti-export bias (as in Appendix 2) was used to apportion the share of manufacturing output subjected to an anti-export bias in Appendix 3. However, in terms of the classifications reflected in Figures 2 and 3 the sectors subjected to an anti-export bias accounted for 5 per cent of total output in 1990 and 21 per cent in 1999. Thus, while more sectors were subjected to an anti-export bias in trade policy during the end of the 1990s than in the earlier part of the decade, the extent of anti-export bias prevailing during this period was not as high as has been claimed in the empirical literature (see Belli et al., 1993; IDC, 1996a; Tsikata, 1999). The results are not much different if the analysis is confined to the manufacturing sector. The share of manufacturing output subjected to an anti-export trade policy bias was 7 per cent in 1990 and 32 per cent in 1999. In these calculations the ‘other mining’ sector was also included in the anti-export bias group. Technically speaking, this sector should not be included in this group – as argued earlier on, in a conventional sense anti-export bias involves a disincentive for export production but an incentive for import substitution. The results are thus biased towards an increase in the group comprising those sectors subjected to an anti-export bias trade regime.

4. Trade incentives and output growth in the South African economy during the 1990s

This section considers how output growth of the different sectors related to the trade policy bias identified in the previous section. All things being equal, and under the conditions of perfect competition, the trade policy and output effects of the five trade policy regimes (identified in Figure 1) can be summarised as in Table 2. For example, under an import substituting regime there is a positive trade incentive effect on import substitution but a negative trade incentive effect on export promotion. Hence under import substitution there is a positive impact on import substituting production but a negative impact on export production. As pointed out earlier, an anti-export bias exists when there is a trade policy bias against export production. This is only the case with the DIP and IS trade strategies.

Table 2:  Trade regime effects on output and trade policy

Trade regime	Trade incentives effect		Output effects
	Import substitution	Export promotion	Import substitution	Export production
IS	+ve	−ve	+ve	−ve
PEP	+ve	+ve	+ve	+ve
EP	−ve	+ve	−ve	+ve
DIP	−ve	−ve	−ve	−ve
Free trade
Source: Author's deductions.

Following Chenery (1979), changes in gross production can be allocated across the demand factors of domestic demand, export expansion and import substitution as follows:

where: Q = gross value of output, DD = domestic demand, EE = export expansion, IS = import substitution

Equation 8 may be reformulated as follows:

where m _t , Q _t , E _t represents the import coefficient (defined as the share of imports in the domestic demand), output and exports in year t.

The first term on the right-hand side is the contribution of domestic demand to the growth of gross output. The second term captures the effect of export expansion. The third term reflects the change in the import coefficient for a given level of domestic demand; a positive sign indicates that import substitution has taken place, while a negative sign means that foreign goods have gained market share.

Equation 9 is used to calculate the sources of growth for the various sectors during the period under analysis. Table 3 captures the bias in the trade regime (columns 3 and 7) and the source of growth (columns 4, 5, 6, 8, 9, 10) for the different sub-periods under analysis. From columns 3 and 7 it is evident that there were 20 sectors that had some bias in their trade regime during the 1990s – this information was derived from Figures 2 and 3.

Table 3:  Trade regime bias, production bias and sources of economic growth

	Sector [Standard industrial classification]	1990–1994 Sources of growth				1995–1997 Sources of growth				1995–2001 Sources of growth
		TRB^1	DD	EE	IS	TRB^1	DD	EE	IS	DD	EE	IS
[1]	[2]	[3]	[4]	[5]	[6]	[7]	[8]	[9]	[10]	[11]	[12]	[13]
1	Agriculture, forestry and fishing		1.78	0.23	−1.01	IS	1.35	−0.10	−0.25	0.67	0.54	−0.21
2	Coal mining		0.55	0.46	−0.01		0.52	0.52	−0.04	−0.37	1.48	−0.11
3	Gold and uranium ore mining		0.00	1.00	0.00		0.00	1.00	0.00	0.00	1.00	0.00
4	Other mining	DIP	−0.05	0.68	0.37	DIP	0.53	0.19	0.29	0.26	0.63	0.11
5	Food	PEP	0.98	0.06	−0.04	PEP	0.98	0.13	−0.11	0.86	0.18	−0.04
6	Beverages		0.56	0.53	−0.09		0.80	0.29	−0.09	1.11	−0.04	−0.07
7	Tobacco	IS	0.89	0.11	0.00	IS	0.55	0.48	−0.03	0.61	0.44	−0.04
8	Textiles	IS	0.99	0.07	−0.07	IS	0.67	0.44	−0.11	0.98	0.06	−0.03
9	Wearing apparel		0.79	0.30	−0.08	IS	0.77	−0.05	0.28	0.43	0.24	0.33
10	Leather and leather products	PEP	0.62	0.73	−0.35	PEP	0.57	0.56	−0.13	0.86	0.20	−0.06
11	Footwear	PEP	0.76	−0.12	0.36	IS	0.87	−0.12	0.25	0.17	−0.06	0.88
12	Wood and wood products		0.92	0.16	−0.08		0.92	0.08	0.00	1.39	−0.35	−0.05
13	Paper and paper products		0.97	0.14	−0.11		−1.42	2.46	−0.04	−0.25	1.27	−0.02
14	Printing, publishing and recorded media	PEP	0.98	0.02	0.00	IS	1.12	0.03	−0.15	1.10	0.14	−0.24
15	Coke and refined petroleum products		0.88	0.09	0.03		0.74	0.43	−0.16	−0.24	1.00	0.24
16	Basic chemicals		1.06	−0.24	0.18		0.48	0.52	0.00	0.59	0.41	0.00
17	Other chemicals and man-made fibres		0.99	0.10	−0.09		0.95	0.24	−0.18	1.01	0.38	−0.39
18	Rubber products	PEP	1.01	−0.03	0.03	PEP	0.85	0.85	−0.70	0.65	1.21	−0.86
19	Plastic products	PEP	0.98	0.05	−0.03	IS	0.96	0.17	−0.13	1.19	0.42	−0.61
20	Glass and glass products	PEP	0.99	0.10	−0.09	IS	0.05	−0.87	1.82	0.67	−0.59	0.93
21	Non-metallic minerals		0.94	0.21	−0.16		1.00	0.10	−0.10	1.14	0.50	−0.64
22	Basic iron and steel	PEP	0.37	0.63	0.00	PEP	0.43	0.66	−0.09	0.63	0.40	−0.02
23	Basic non-ferrous metals	IS	1.63	0.07	−0.70		0.64	0.39	−0.04	−0.12	1.41	−0.29
24	Metal products excluding machinery	PEP	0.90	0.12	−0.02		0.82	0.22	−0.04	−0.01	0.41	0.60
25	Machinery and equipment		1.65	0.52	−1.17		0.80	0.67	−0.48	−0.14	−0.65	1.79
26	Electrical machinery and apparatus		1.01	0.16	−0.17	IS	0.25	0.37	0.38	0.23	−0.12	0.89
27	Television, radio and communication equipment	PEP	4.85	0.64	−4.49		−0.53	−0.60	2.14	0.52	0.91	−0.43
28	Professional and scientific equipment	IS	0.76	0.38	−0.14	PEP	1.16	1.67	−1.83	0.62	0.94	−0.57
29	Motor vehicles, parts and accessories		0.55	0.29	0.16	IS	−0.02	0.85	0.17	0.35	−0.26	0.91
30	Other transport equipment		1.82	0.96	−1.77		−0.82	0.78	1.04	−4.85	5.30	0.56
31	Furniture	PEP	1.03	−0.69	0.67		0.39	0.70	−0.09	−0.04	1.36	−0.32
32	Other manufacturing		1.59	2.12	−2.71		0.44	0.55	0.01	0.96	1.58	−1.54
^1Trade regime bias. No entry against any sector in this column implies that there is no bias in the trade regime (i.e. free trade stance).
Source: Own calculations with data from TIPS.

Considering first the period 1990–94, there were five sectors (other mining; tobacco; textiles; basic non-ferrous metals; professional and scientific equipment) that had a trade incentive bias against export production. Of interest, however, is that export production still accounted for some of the output growth in all these sectors. The picture does not change much for the period 1995–97. Of the 12 sectors (agriculture, forestry and fishing; other mining; tobacco; textiles; wearing apparel; footwear; paper and paper products; printing, publishing and recorded media; plastic products; glass and glass products; electrical machinery and apparatus; motor vehicle parts and accessories), four sectors (agriculture, forestry and fishing; wearing apparel; footwear; glass and glass products) showed negative contributions from export production to output growth. Export production made a positive contribution to output growth in the remaining eight sectors.

A conclusion to be drawn from Table 3 – albeit a tentative one – is that in general there was not a strong correlation between the trade policy stance that sectors were subjected to and production outcomes. As pointed out in section 2, trade policy is one element influencing production outcomes. Any appraisal has to consider how trade policy related to the other determinants of production. South Africa has engaged in extensive tariff liberalisation during the 1990s (see Rangasamy & Harmse, 2003) and there are no policy intentions to reverse this trend. Given the tariff reductions that have been implemented since 1997 under the auspices of the WTO, the anti-export bias in the trade regime would in all probability have decreased further relative to the period 1995–97. The issue may therefore be that it is not tariff protection per se but other factors related to domestic competitiveness (e.g. skills development, productivity enhancement, competition policy, exchange rate, etc.) and market access that pose the major challenges to export production in South Africa. This issue warrants further empirical research.

5. Concluding remarks

An analysis of trade incentives in the South African economy during the 1990s reveals two major findings. Firstly, the extent of the anti-export bias in South Africa's trade policy during the 1990s is lower than that claimed in the empirical literature. The sectors subjected to an anti-export bias accounted for around 7 per cent of total output in 1990 and 21 per cent in 1999. Secondly, sectoral output growth did not strongly correlate with the prevailing trade incentives of the 1990s. Export production continued despite the prevalence of import substituting incentives in many sectors. Although further research is needed in this regard, this suggests that attention should also be given to other factors related to domestic competitiveness (e.g. skills development, productivity enhancement, competition policy, etc.) and market access if South African export production is to be significantly increased.

The views expressed in this article do not necessarily reflect those of the South African Reserve Bank. I am grateful to Greg Farrell (SARB) and an anonymous referee for comments on an earlier draft. Any errors are, of course, my own.

Appendix 1

Anti‐export bias in the South African economy

	SIC^1	Industry	1993	1999
1	1100	Agriculture, forestry and fishing	1.08	1.28
2	2100	Mining	1.03	1.03
3	3111	Slaughtering preparing & preserving meat	20.58	30
4	3112	Diary products	1.05	8.46
5	3113	Canning & preserving of fruit & vegetables	1.04	2.29
6	3114	Canning preserving & processing of fish	1.03	0.87
7	3115	Vegetable & animal oils & fats	2.08	1.15
8	3116	Grain mill products	0.98	0.81
9	3117	Bakery products	1.51	7.15
10	3118	Sugar factories & refineries	1.12	30
11	3119	Cocoa, chocolate & sugar confectionery	0.98	1.30
12	3121	Other food products	3.38	1.24
13	3122	Prepared animal feeds	0.48	5.22
14	3131	Distilleries & wineries	0.86	8.70
15	3133	Malt liquors & malt	0.81	0.94
16	3134	Soft drinks & carbonated waters industries	1.42	1.04
17	3140	Tobacco products	7.89	30
18	3211	Spinning, wool weaving & finishing of fabrics	3.76	2.60
19	3212	Made-up textile goods, excl. wearing apparel	1.74	3.39
20	3213	Knitting mills	6.27	1.98
21	3214	Carpets & rugs mats & matting	1.30	3.56
22	3215	Cordage, rope & twine industries	0.59	4.60
23	3219	Textiles, not elsewhere classified	0.35	1.46
24	3220	Wearing apparel excl. footwear	7.91	7.07
25	3233	Leather products & leather substitutes	1.33	3.36
26	3240	Footwear	1.78	2.26
27	3310	Wood & wood products excl. furniture	1.12	1.06
28	3320	Furniture	1.22	2.39
29	3411	Pulp, paper & paperboard	1.00	1.18
30	3412	Paper containers	1.09	1.66
31	3419	Other pulp, paper & paperboard	1.09	1.59
32	3420	Printing & publishing	0.79	0.97
33	3511	Industrial chemicals	0.93	1.02
34	3512	fertilisers & pesticides	0.77	1.00
35	3513	Synthetic resins & plastic materials	1.26	1.16
36	3521	Paints vanishes & lacquers	2.53	1.23
37	3522	Medical & pharmaceutical preparations	0.85	0.96
38	3523	Soap, cosmetics & toilet preparations	2.13	2.52
39	3529	Other chemical products	1.13	1.02
40	353/4	Other basic chemicals petroleum & coal	1.87	1.82
41	3551	Tyres & tubes	1.84	1.75
42	3559	Other rubber products	1.34	1.28
43	3560	Other plastic products	1.18	1.84
44	3610	Pottery, china & earthenware	1.21	1.77
45	3620	Glass & glass products	1.18	1.19
46	3691	Bricks, tiles, refractories, etc.	1.20	1.39
47	3692	Cement	0.86	0.97
48	3699	Other non-metallic mineral products	1.41	1.23
49	3710	Iron & steel basic industries	1.12	1.08
50	3720	Non-ferrous metal basic industries	1.26	1.07
51	3811	Cutlery, hand tools & general hardware	1.03	1.63
52	3312	Furniture & fixtures primarily of metal	1.11	1.89
53	3813	Structural metal products	0.85	1.22
54	3819	Other fabricated metals excl. machinery	1.11	1.29
55	3821	Engines & turbines	0.74	1.07
56	3822	Agriculture machinery & equipment	0.57	0.95
57	3823	Metal & woodworking machinery	0.68	0.99
58	3824	Special industrial machinery & equipment	0.67	0.97
59	3825	Office computing & accounting machinery	0.54	0.96
60	3829	Other machinery & equipment excl. electrical	0.83	1.01
61	3831	Electrical industrial machinery	0.94	1.09
62	3832	Radio, television & communication equipment	0.95	1.10
63	3833	Electrical appliances & housewares	1.08	2.49
64	3839	Other electrical apparatus & supplies	0.99	1.05
65	3843/0	Motor vehicles	3.66	4.68
66	3852	Railway equipment	0.66	0.96
67	3851/4	Other transport	0.92	1.06
68	3901	Jewellery and related articles	0.48	0.97
69	386,390	Other manufacturing industries	1.05	0.92
70	41	Services	0.95	0.97
		Total Industry	1.19	1.32
		Total Manufacturing	1.27	1.44
Notes: ^1Standard industrial classification.
Source: IDC, 1996a: The effective protection rate and anti-export bias, IDC, Sandton.

Appendix 2

Anti-export bias in the manufacturing sector based on value of sales

SIC^1	Sector description	Anti-export bias		Value of sales (Rm)
		1993	1999	1990	1999
3111	Slaughtering, processing and preserving of meat	X	X	2 653.49	5 693.37
3114	Processing & preserving of fish & fish products	X		794.37	1 533.32
3114	Processing & preserving of fruit & vegetables	X		2 179.75	4 489.28
3115	Vegetable & animal oils & fats	X	X	1 968.84	5 566.10
3112	Dairy products	X	X	3 226.55	7 164.65
3116	Grain mill products			5 003.86	9 847.01
3122	Prepared animal feeds		X	2 379.89	5 024.87
3117	Bakery products	X	X	2 598.22	4 156.73
3118	Sugar refining	X	X	2 124.42	4 473.10
3119	Cocoa, chocolate & sugar confectionery		X	1 204.45	2 910.92
3121	Other food products	X	X	2 580.46	6 419.42
3131	Distilling industries		X	2 698.89	5 545.56
3133	Beer, malt liquors & malt			4 235.16	8 967.19
3134	Soft drinks & mineral waters	X	X	2 266.46	6 092.40
3211	Spinning & weaving of textiles	X	X	3 906.72	6 304.57
3212	Made-up textile articles excl. apparel	X	X	991.78	1659.44
3214	Carpets, rugs and matting	X	X	458.95	701.13
3219	Other textiles		X	422.93	1 105.15
3213	Knitted & crocheted fabrics & articles	X	X	1 195.06	1 858.59
3220	Wearing apparel excl. fur apparel	X	X	4 848.05	9 135.95
3233	Tanning and dressing of leather	X	X	871.35	2 577.53
3240	Footwear manufacturing	X	X	1 806.70	2 379.13
3310	Sawmilling and planing of wood	X	X	1 362.29	2 380.67
3319	Wood and wood products	X	X	2 091.36	6 005.76
3411	Pulp, paper and paperboard	X	X	4 360.68	10 019.29
3412	Corrugated paper & paperboard	X	X	2 970.35	6 814.00
3419	Other articles of paper & paperboard	X	X	1 184.54	3 567.30
3420	Publishing			1 904.56	4 471.08
3420	Printing and related services			3 201.96	7 193.86
3540	Coke oven products	X	X	455.98	900.15
3540	Petroleum refineries & synthesisers	X	X	9 938.68	20 099.54
3511	Basic chemicals		X	3 233.44	6 881.96
3512	Fertilisers and nitrogen compounds		X	1 790.11	5 932.42
3513	Plastics in primary form	X	X	3 218.18	7 058.53
3511	Pesticides and other agro-chemical products		X	712.14	1 391.37
3521	Paints, varnishes, printing ink and mastics	X	X	1 252.86	3 683.35
3522	Pharmaceuticals, medicinal chemicals, etc.			2 795.58	5 781.05
3523	Soap, detergents & polishing, perfumes, etc.	X	X	2 519.25	7 605.54
3529	Other chemical products	X	X	3 563.37	7 028.73
3551	Tyres & tubes of rubber	X	X	1 743.79	3 194.88
3559	Other rubber products	X	X	875.16	972.35
3220	Plastic products	X	X	4 327.88	9 723.28
3620	Glass and glass products	X	X	1 612.06	2 140.82
3610	Ceramic Non-structural non-refractory	X	X	225.51	269.80
3691	Ceramic products (refractory)	X	X	1 465.89	2 431.05
3692	Cement, lime and plaster			1 224.43	2 313.10
3699	Concrete, cement or plaster articles	X	X	2 604.93	4 243.19
3710	Basic iron and steel	X	X	13 362.06	26 056.85
3720	Basic precious and non-ferrous metals	X	X	4 447.31	13 107.16
3813	Structural metal products and steam generators		X	4 581.48	8 375.92
3819	Forging, pressing, stamping, etc. of metal	X	X	1 254.54	2 181.45
3811	Cutlery, hand tools & general hardware	X	X	974.88	1 839.26
3829	Other fabricated metal products		X	5 962.54	11 054.27
3829	General purpose machinery and computing machinery		X	5 206.56	8 715.99
3829	Special purpose machinery		X	5 268.12	7 840.00
3829	Household appliances		X	1 204.89	1 926.20
3831	Electric motors, generators & transformers		X	654.82	1 736.56
3839	Electricity distribution & controlling apparatus		X	629.52	2 129.98
3839	Insulated wire & cable manufacturing		X	1 510.57	2 683.71
3839	Accumulators, primary cells & primary batteries		X	582.81	1 004.13
3839	Electric lamps & lighting equipment		X	490.84	771.88
3831	Other electrical equipment		X	1 951.93	3 509.46
3832	Radio, television and communication		X	2 570.88	4 441.40
3829	Medical equipment, instruments		X	602.03	1 299.95
3840	Motor vehicles	X	X	10 732.67	30 465.94
3840	Bodies for motor vehicles, trailers & semi-trailers	X	X	1 058.11	1 529.08
3840	Motor vehicle parts and accessories	X	X	3 705.99	10 724.81
3840	Other transport equipment	X	X	1 539.35	2 789.06
3320	Furniture manufacturing	X	X	3 027.74	6 356.28
3901	Jewellery and related articles			1 558.16	2 449.12
3909	Other manufacturing industries	X		3 988.94	9 991.86
	Total manufacturing sales			187 918.07	398 688.75
	% of manuf. sales subject to anti-export bias			0.66	0.86
Notes: ^1Standard industrial classification.
Source: IDC 1996a.

Appendix 3

Total output subject to anti-export bias (based on 2 sectors classification)

Sector description	Output
	1990	1999
1. Agriculture, forestry and fishing (SIC 1100)	19 818.8	49 431
2. Mining (SIC 2100)	42 976.8	86 972
3. Services (SIC 4100)	281 802	843 123
4. Manufacturing^1(SIC 3000)	135 909	388 055
5. Total output	550 521	1 430 753.5
Per cent of total output subject to AEB	0.87	0.96
Notes: ^166 per cent (86 per cent) of the output of the manufacturing sector is assumed to be produced under an anti-export bias trade regime. See Appendix 2 for calculations.
^2Given by [(1 + 2 + 3 + 4)/5].
Source: IDC, TIPS.

Appendix 4

Total output subject to anti-export bias (based on 3 sectors classification)

1. Sectors with anti-export bias (1990-94) Rm	1990
1.1 Other mining	15 592.4
1.2 Textiles	6 345.5
1.3 Tobacco	1 900.6
1.4 Basic non-ferrous metals	4 786.7
1.5 Professional and scientific equipment [374–376]	668.4
2. Manufacturing output	205 923.1
3. Total output (overall economy)	550 521.3
4. Per cent of total output with AEB^1	0.05
5. Per cent of manf output with AEB^2	0.07
6. Sectors with anti-export bias (1995–97) Rm	1999
6.1 Agriculture, forestry and fishing	49 431
6.2 Other mining	40 865.1
6.3 Tobacco	4 943.3
6.4 Textiles	11 076.3
6.5 Wearing apparel	12 324.5
6.6 Footwear	2 708.8
6.7 Paper and paper products	22 636.4
6.8 Printing, publishing and recorded media	13 265.5
6.9 Plastic products	11 072.8
6.10 Glass and glass products	2 415.6
6.11 Electrical machinery and apparatus	13 417.3
6.12 Motor vehicles, parts and accessories	48 619.6
7. Manufacturing output	451 226.9
8. Total output (overall economy)	1 121 179
9. Per cent of total output with AEB^3	0.21
10. Per cent of manuf. output with AEB^4	0.32
Notes: ^1Given by [(1.1 + 1.2 + 1.3 + 1.4 + 1.5)/3].
^2Given by [(1.2 + 1.3 + 1.4 + 1.5)/2].
^3Given by [(6.1 + 6.2 + 6.3 + 6.4 + 6.5 + 6.6 + 6.7 + 6.8 + 6.9 + 6.10 + 6.11)/8].
^4Given by [(6.3 + 6.4 + 6.5 + 6.6 + 6.7 + 6.8 + 6.9 + 6.10 + 6.11)/7].