ARTICLE SOUTH AFRICAN REGIONAL INDUSTRIAL DEVELOPMENT POLICY: CRITICAL ISSUES Trevor Bell 1. Introduction The aim of this paper is to set out the central issues for regional industrial development policy in South Africa today.1 Section 2 considers the usual international objectives of regional policy. Section 3 outlines the view of the Report of the Panel of Experts (POE, 1989), which underlay the Regional Industrial Development Programme (RIDP), introduced in 1991, that the primary aim of regional policy in South Africa should be economic, and that the targeting of centres in and near the former homelands, as in the 1982-91 RIDP, was neither compatible with this aim, nor any longer necessary on social grounds. Section 4 considers the view that a regional industrial development policy which concentrates particularly on the poorer regions of South Africa, for social or political reasons, is now superfluous. Section 5 asks whether, as claimed by the POE (1989), the 1982-91 REDP's system of preferential incentives for centres in and near the former homelands was indeed in conflict with the promotion of 'the self-sustaining economic growth and development' of the South African economy as a whole. In particular, under various sub-headings, it questions the POE's view that the system of nationally more uniform incentives introduced in 1991 is economically optimal, and that the centres especially favoured under the 1982-91 RIDP lacked a comparative advantage in manufacturing activity. It is suggested that a regional industrial development policy which targets such centres has a useful role to play in promoting structural adjustment, particularly in the context of South Africa's trade policy reforms. Section 6 argues that, the POE's views notwithstanding, the targeting of low-wage areas in 1982-91 was in fact most beneficial to the people and to overall economic development in those areas, and can continue to be so in future. Section 7 is concerned with the question whether the availability of agriculture as a possible alternative, undermines the case for the promotion of manufacturing industry in the poorer regions of South Africa, or whether the roles of agriculture and industry in regional development are essentially complementary. Section 8 considers the reasons for the policy shifts of 1982 and 1991, and for the direction TRANSFORMATION 32 (1997) ARTICLE BELL which regional industrial development policy has taken in 1996, in terms of swings in socioeconomic conditions and in the prevailing economic and social philosophy. Section 9 discusses the policy implications of the foregoing sections, while taking into account some additional matters not considered earlier in the paper. The discussion ends with a postscript consisting of a description of, and some comments on, the main features of the policy introduced in 1996, since the writing of the initial version of the paper. 2. The Objectives of Regional Policy The objectives of regional policy vary from one country to another, and, indeed, also, within any particular country, from one period to another. Generally, however, regional policy aims to improve economic conditions in problem areas. It thus generally seeks to contribute to areduction in inter-regional income inequalities, as well as in the unemployment rate and the incidence of poverty, in areas where these are most severe. Regional policy may also be directed at promoting structural adjustment, and realising a country's potential for economic development more fully (Hansen et al, 1990:282). Regional development policies, though, are generally not motivated primarily by considerations of economic growth and efficiency. This clearly has also been the case in South Africa, during most of the history of regional industrial development policy in this country. From the beginning of systematic attempts to influence the geographical distribution of manufacturing industry, with the announcement of the 'border industries' policy in June 1960, to the advent of the 1991 RIDP, South African regional industrial development policy was primarily an instrument of grand apartheid. Given this, the main focus of the development effort was on the promotion of manufacturing activity in or near the black 'homelands'. In terms of per capita income levels, unemployment rates and the incidence of poverty, the black homelands were also the most economically deprived regions of South Africa. Clearly, in targeting centres in and near these territories, South African regional industrial development policy was not motivated by considerations of inter-regional equity, let alone social justice. The fact remains however that, as in many other countries, regional industrial development policy in South Africa did concentrate its efforts on the promotion of manufacturing activity in the country's poorest regions. The 1991 RIDP represented a radical break with the past. Its ostensible aim is apparently primarily economic, in particular, promotion of 'the self-sustaining economic growth and development of the integrated [South African] economy' (POE, 1989:xxix). It introduced a system of incentives applicable more or less TRANSFORMATION 32 (1997) LL ARTICLE iformly to manufacturing plants in all centres outside the PWV, urban-Pinetown and Cape Town metropolitan areas. Plants situated in or near e homelands thus ceased to be eligible for especially favourable financial ducements. Premises of the 1991 RIDP There are apparently three ground* on which the drafters of the 1991 RIDP andoned the principle of significant preferential treatment for centres in or near e homelands. One reason given is simply that they regarded considerations her than economic efficiency as lying outside their terms of reference (POE, 89:290). Thus, they dismiss POE, (1989:290), as 'political considerations', e arguments put by the authors of the minority report that 'to ensure stability, ealth must also be created in backward areas', and that the proposed new policy ould be a 'step backwards from the concept of redressing the economic balance between White and Black areas' (POE, 1989:265). These are regarded considerations which it was 'agreed at the outset were not part of their brief OE, 1989:290). In fact, however, the interpretation of their terms of reference in this manner as based on two substantive arguments for abolishing the principle of especially vourable treatment for centres in or near the homelands. One of these arguments is implicit in such statements, that'... a decentralisation rategy needs to be targeted at localities where there is the highest probability at the conditions exist or can easily be realised for new investment to flourish a manner that leads to further business opportunities as well as the capacity to ploit them. The targets clearly must be the country' s secondary cities, not small wns or isolated locations' (POE, 1989:106); that 'the differentiation in the isting incentives are such that locations with the greatest natural development tential tend to receive the smallest incentives' (POE, 1989:117); and that: olicy will be most successful when based on proven economic principles [such ] comparative advantage ... [for then] policy would assist market forces and t waste resources by trying to reverse them' (POE, 1989:291). The current RIDP thus is based on the view that the centres in and near the melands did not have a comparative advantage in manufacturing activity; that ey therefore had little potential for sustainable industrial development; and that ecial incentives designed to induce industrial growth in such centres presented an economically wasteful attempt to swim against the tide of market rces. In short, there was no economic case for special inducements, and the se for promotion of industrial development in these centres rested solely on RANSFORMATION 32 (1997) 3 ARTICLE BELL social objectives, which clashed head-on with the country's objective of faster economic growth. The second substantive argument is in effect that the abolition of legal obstacles to rural-urban migration had rendered unnecessary the pursuit of industrial development in and near the homelands, for social or political reasons. In response to the argument of the minority report that such development was intended to compensate for the restrictions on the movement and settlement of blacks, it is stated that 'cognisance is taken of relaxations in the restrictions on Blacks that have already taken place in South Africa', and that 'the recommendations are made on the assumption that this trend will continue' (POE, 1989:290). In essence, the present RIDP was based on the premises that the targeting of centres in and near the homelands was neither compatible with its declared overall goal of promoting 'the self-sustaining economic growth and development of the integrated [South African] economy' (POE, 1989:xxix), nor necessary on social grounds, to address inter-regional inequalities. These assertions raise issues which are central to the evaluation of the RIDP, and are the main focus of the discussion below. 4. Inter-Regional Inequalities and the Need for a Regional Development Policy Consider first the view that it is no longer necessary on social or political grounds for regional development policy to concentrate particularly on the poorer regions of South Africa. One of the features of developing countries, distinguishing them from advanced industrial countries, is the generally greater size of their inter-regional income inequalities. In this respect South Africa is clearly no exception. A recent report by the Centre for Development and Enterprise (CDE, 1995:19) finds that in 1995 '41 percent of the South African black population lived in metropolitan areas and received 64 percent of [black] personal income. Another 18 percent lived in cities/large towns and received 18 percent of [black] personal income [and] 41 percent lived in small towns and received only 18 percent of [black] personal income'. This seems to imply that the per capita incomes of blacks in small towns and remote rural areas were only about 28 percent of those in metropolitan areas and 44 percent of those in cities/large towns. Other information provided in the same report also seems to imply that the proportion of people living in poverty in small towns and remote rural areas was three times greater than in the rest of the country. TRANSFORMATION 32 (1997) BELL ARTICLE Furthermore, though it is estimated that the proportion of South Africa's population living in rural areas will decline (from 43 percent in 1995 to 42 percent in the year 2000 and 38 percent in 2011), even 'by 2011, as many South Africans will be living in rural and small town contexts as those who live in the metropolitan areas' (CDE, 1995:22). Given the size of the population of small towns and rural areas, and the disparity between the per capita incomes and the incidence of poverty in those areas compared to the rest of the country, South Africa clearly has a major regional problem, which cannot be ignored. Does the abolition of influx controls, and hence of legal obstacles to migration from small towns and rural areas, however, render redundant regional development policy aimed at redressing inter-regional inequalities? The CDE study finds (1995:9) that current expectations of the growth rate of South Africa's total population, and in particular of the population of the metropolitan areas, are significantly lower than was commonly expected a few years ago. They now estimate that by the year 2000 there will be five million fewer people in the metropolitan areas than was previously expected, and that by 2010 there will be 8.5 million less. One possible interpretation of this is that following a period of rapid rural-urban migration, the rate of migration has slowed, due to a new 'dynamic equilibrium' between rural and urban areas, reflecting the voluntary choices of people in small towns and rural areas. The CDE report states (1995:22): 'During the 1980s in general, and in the first half in particular, there was a very clear imbalance between conditions in rural and urban areas which manifested itself first in the defacto and later in the dejure breakdown of influx control, and rapid, unplanned urbanisation'. It adds that 'following the collapse of that system during the 1970s and 1980s, and the attendant flow of people to the cities, it appears that the pattern of costs and benefits — is such that metropolitan destinations 'are not necessarily the automatic choice for all migrants'. Evidence for the emergence of such a new dynamic equilibrium, however, is lacking. Indeed, it is uncertain even whether there has in fact been a slow-down in the rate of rural-urban migration. As the CDE (1995:28) notes, its 'demographic model does not deal with why the population of the metropolitan areas is growing at a slower pace', and ' ... a large part of the reason for the relatively diminished growth of the metropoles has to do with lowered rates of [natural] population increase within the metropoles themselves' (1995:22). Hence they say: 'It would... be an error to conclude that migration to these areas is slowing down. It might be but we do not know - we can only speculate'. TRANSFORMATION 32 (1997) ARTICLE . BELL In the final analysis, however, neither evidence for a slowdown nor an increase in the rate of rural-urban migration on its own has unambiguous implications for regional development policy. A slowdown may well be due to negative developments in the metropolitan areas (in particular, for instance, as the CDE report (1995:23) suggests, to 'the metropoles not generating economic opportunities on scale and the violence associated with the large metropolitan regions'), rather than to any improvement in the situation of rural relative to urban people. Nor would evidence of significant continuing migration necessarily mean that migration was tending to produce a convergence of living standards inter-regionally, and that the regional problem, as characterised above, was being overcome in this way. Much depends on the forces underlying the migration. For instance, the rural areas may continue to disgorge people because of harsh and perhaps deteriorating rural and agricultural conditions, and rapid rural-urban migration thus will not necessarily be accompanied by an improvement in conditions in small towns and rural areas relative to the cities and metropolitan areas. Both a slowdown and acceleration of migration from small towns and rural areas, thus, are compatible with the persistence of large inter-regional inequalities. They are also compatible with a large body of poor being left behind in those areas, unable to improve their lot through migration. One of the implications of the above is that we badly need a much better understanding of the rate and determinants of rural-urban migration. In the present state of knowledge we are left, however, with the fairly sure conclusion that there are very large inter-regional disparities in per capita incomes and in the incidence of poverty, and that migration cannot be depended upon to reduce them, even in the absence of legal obstacles to migration. If migration could be depended on to perform this function, there would be little motivation anywhere for governmental programmes aimed at developing problem regions. Such programmes, however, have long existed in many other countries, where, like South Africa today, there have been no legal obstacles to inter-regional migration. There are many barriers to migration other than legal ones. Contrary to what seems to be one of the premises of the current RIDP, the need for a regional development which targets poor regions, on grounds of equity or social justice, is not eliminated by the removal of influx controls. Indeed, with the end of apartheid, which gave regional policy in South Africa a bad name, the case on these grounds for a regional development policy is probably clearer than ever. Though referring specifically to industrial development, the following remarks written some years ago now (Bell, 1987:217-8), seem relevant to regional development policy in general today: TRANSFORMATION 32 (1997) BELL ARTICLE It is arguable that, without influx control, the case for decentralisation policy would be greater than ever before. Instead of merely aiming to alleviate the effects of influx controls, and in the process possibly tending to justify them, decentralisation policy could be used for the more legitimate task of improving the living conditions of the genuinely immobile poor in the least developed parts of the country. Under present conditions there is a large body of unemployed, even in the majormetropolitan areas; but the largest element of those in extreme poverty is in the homelands. Thus far we have argued only that South Africa has a major regional problem, and that there is a need on grounds of equity and social justice for economic development of the poorer regions of South Africa. There remains the crucial question of whether, and to what extent, a programme of regional development, and in particular of regional industrial development, is economically feasible. 5. Is there an Economic Case for Targeting Centres in and near the Former Homelands? As noted above, the current RIDP's ostensible aim is the promotion of 'the self-sustaining economic growth and development of the integrated [South African] economy' (POE, 1989:xxix). Underlying the RIDP is the view that the targeting of centres in and near the former homelands is not compatible with this overall aim, and that the targets of regional policy must be the country's 'secondary cities', on the grounds that these have the greatest natural development potential. It is based on the view that centres in and near the homelands did not have a comparative advantage in manufacturing activity, that they therefore had little potential for sustainable industrial development, and that special incentives designed to induce industrial growth in such centres represented an economically wasteful attempt to swim against the tide of market forces. It therefore offers a system of incentives applicable more or less uniformly to manufacturing plants in all centres outside of the PWV, Durban-Pinetown and Cape Town metropolitan areas. The rest of this section considers the issues arising from these features of the current RIDP, and suggests some alternative perspectives, with different implications for regional industrial development policy. Is a Uniform System of Incentives Economically Optimal? The provision of financial incentives to plants establishing in centres outside the country's majormetropolitan areas, was justified by the drafters of the current TRANSFORMATION 32 (1997) ARTICLE BELL RIDP on two theoretical grounds (POE, 1989:ch7). One is that the growth of the PWV region had been economically excessive, taking into account the size of the external diseconomies associated with metropolitan expansion, relative to positive agglomeration economies, as well as subsidiary factors such 'locational inertia', lack of full information required for the selection of profit-maximising locations, the costliness of relocation, and the unpredictability of future changes in spatial costs and prices. The second part of the theoretical argument involved references to divergences between private and social costs of production in centres situated outside the PWV region, which it was contended give such plants 'infant industry status'. Financial incentives for plants outside the PWV area thus would simultaneously serve to curb further economically excessive expansion of the PWV region, and provide support, justifiable on infant industry grounds, for plants outside of this region. The case for the geographical disconcentration of industry, away from the PWV on the grounds of the external diseconomies of agglomeration is questionable, and, furthermore, it is doubtful that the system of financial incentives introduced in 1991 tends to correct for such externalities. For the sake of argument, however, let us accept that there is an economic case on such grounds for curbing the further growth of the major metropolitan areas, by providing financial incentives at centres outside these areas. Is a system of subsidies applicable more or less uniformly to such centres economically optimal? As noted above, apart from the external diseconomies generated by expansion in the metropolitan areas, the case for incentives outside these areas depends, according to the drafters of the current RIDP, on divergences between private and social costs associated with the infant industry argument. It is highly unlikely that a uniform system of incentives, aimed at correcting for such divergences, will achieve the stated goal of the policy, 'to create an environment in which the decisions of all industrialists are based on the true or social costs of production' (POE, 1989:165). It would do so only if the size of the divergences was uniform across all centres outside the major metropolitan areas. It is virtually certain that this condition for a uniform system of subsidies to be economically optimal does not hold. It is rather less certain how the divergences between private and social costs associated with the infant industry argument, which apparently underlie the current RIDP, would vary from one centre outside the metropolitan areas to another. However, it seems not unreasonable to expect that such divergences, as well as those associated with an 'infant areas' argument, would in general be greater in relatively new, TRANSFORMATION 32 (1997) LL ARTICLE enfields' sites, in and near the former homelands, than in secondary cities. so far as regional industrial development policy is to be based on such siderations, thus, there would seem to be aprimafacie case for a system of ucements which is not uniform across centres outside the metropolitan areas, which gives preferential treatment to existing, or potential, newly ustrialising areas.2 Centres in and near the Former Homelands Lack Comparative antage in Manufacturing Activity? has been argued so far only that there may be divergencies between private social cost of the kind associated with the infant industry or infant area uments, and that these may be greater in centres in and near the former elands. The question remains, however, whether these centres have a basic parative advantage in industry, which would enable them to make use of ntives reflecting such divergences. The infant industry argument depends only on the existence of certain kinds of dynamic external economies, but requires inter alia that private costs of production should in due course fall level which would enable the initially protected industries to stand on their feet. As noted earlier, underlying the current RIDP is the belief that centres nd near the former homelands did not qualify for special assistance in terms proven economic principles [such as] comparative advantage ...' (POE, 9:291), or in terms of the criterion that industrial development should go nly to regions 'with a comparative advantage for industry' (POE, 1989:292). y, therefore, did not have significant potential for sustainable industrial elopment. here was substantial growth of manufacturing activity in these centres in the 0s (POE, 1989:25-29). However, in formulating the current RIDP the view taken that 'the decision of the average industrialist to establish development ts was artificial and related to incentives and government intervention, rather to any genuine economic considerations', and therefore that 'a significant ortion of decentralised firms would not be profitable or survive without sidies' (1989:67). The perception was that, in the absence of incentives, stry would in general naturally be drawn to metropolitan areas and secondary s, and that the promotion of manufacturing industry in the industrial elopment points represented an attempt to reverse market forces. hough not explicitly addressed by them, those responsible for the present P implicitly rejected the notion 3 that there had been a spontaneous tendency ards industrial dispersal in South Africa, at least since the early-1970s, and NSFORMATION 32 (1997) ARTICLE BELL that industrial development policy, therefore, had, up to a point at least, been swimming with, rather than against, the tide of market forces. This spontaneous tendency was seen as largely related to a process of structural adjustment in certain traditional, unskilled labour-intensive industries, in the face of intensified international competition, which involved a relative shift in the geographical distribution of such industries towards low-wage areas in and near the former homelands. In short, it was seen largely as a reflection of the comparative advantage of such centres in unskilled labour-intensive industries, based on the availability of abundant supplies of cheap, unskilled labour. According to this view, the special financial incentives may have pushed the process of industrial dispersal further than would have occurred spontaneously, and indeed further than could have been justified on economic grounds, in terms of an infant industry and infant area argument. It may also have had a significant influence on the particular low-wage centres in which the industrial growth actually occurred, making for greater expansion in less advantageously situated development points. There was nevertheless in operation, according to this view, a (probably significant) market-driven tendency for certain labour-intensive industries to shift to centres endowed with abundant supplies of cheap unskilled labour. The case for this view had been based largely on trends in the proportion of manufacturing employment at centres in and near the homelands, and the timing of these trends. In the formulation of the present policy, however, the notion of a significant market-driven tendency towards industrial dispersal was implicitly rejected on the basis of evidence suggesting that most plants in the development points would have been unprofitable without the subsidies. It was stated (POE, 1989:50), for instance: 'A Development Bank of Southern Africa analysis of 55 industries in Ezakheni [Ladysmith], Isithebe and Madadeni [Newcastle] concluded that 42 percent were financially weak, 36 percent financially average and 21 percent viable'. In the light of this and other evidence based on analysis of financial data, they suggest that 'between 40 and 60 percent of decentralised firms are unprofitable and depend on subsidies' (POE, 1989:50). In addition, the findings of various questionnaire surveys are cited, which together 'suggest that a majority (approximately 60 percent) of firms are viable, whereas the evidence from more objective evaluations based on financial statements... points in the opposite direction' (POE, 1989:50). They suggest (on some rather questionable grounds) that the subjective questionnaire surveys tend to exaggerate the profitability of decentralised firms, and that the objective assessments based on financial records tend to underestimate it. They conclude: 'An average figure of 50 percent for the proportion of unprofitable firms 10 TRANSFORMATION 32 (1997) BELL ARTICLE therefore seems to be a reasonable estimate, although this figure is not applicable to all regions and development points' (POE, 1989:52). The important qualifier is inserted that current 'profitability is an inadequate indicator of long-term viability, since the process of development is inherently dynamic' (POE, 1989:52), and because the economy had been in recession since the inception of the 1982 RIDP. Hence, contrary to their earlier estimate, the withdrawal of subsidies would not necessarily result in the closure of all 50 percent of the decentralised firms that were classified as unprofitable. It is nevertheless concluded that the evidence points 'to a lack of success on the part of the programme in promoting viable enterprises' (POE, 1989:54); and 'Far too many decentralised firms are financially weak and dependent on subsidies'. These conclusions based on the evidence cited, appear to have had a major influence in the formulation of the current RIDP. It seems that it was mainly on the basis of this evidence, and the inferences drawn from it, that it was decided that the industrial development points had little comparative advantage in manufacturing activity, and hence little potential for sustainable industrial development, and that special incentives for plants at these development points largely represented a wasteful attempt to swim against the tide of market forces. Unfortunately evidence on the performance since 1991 of plants assisted under the 1982 programme is not available for the industrial development points in general. Such evidence as is available at this stage, however, suggests that the information relied upon in formulating the current RIDP may have exaggerated the non-sustainability of such plants. As noted above, this evidence included a Development Bank of Southern Arica study of plants in Ezakheni (Ladysmith), Isithebe, and Madadeni (Newcastle). In a recent study conducted in Ladysmith and Newcastle, Hart (1995:4) reports as follows: "The sharp decline in RIDP incentives after 1991 has certainly not produced the industrial collapse that many predicted'. Data provided by the KFC for Ezakheni industrial estate suggest that despite some contraction after 1991, there appears to have been remarkable stability in a period of general macroeconomic deterioration in the early-1990s. In Newcastle, according to Hart and Todes (1997), employment in the main manufacturing sectors fell by 21.3 percent between 1990 and 1994. This, however, was mainly due to a decline in employment in heavy industries, particularly ISCOR, which was not established in terms of the 1982 RIDP. Employment in clothing and textiles together in Newcastle fell by 5.5 percent in 1990-94, but this must be seen in the context of an approximately 12 percent decline in employment in these two sectors together in the country as a whole in this period. TRANSFORMATION 32 (1997) 11 ARTICLE BELL Whether experience at Ezakheni and Newcastle is typical of the 1982-91 industrial development points is not known at this stage. It may not be. Indeed, though not presenting supporting evidence, Hart (1995:2) states: 'A major shaking out process ... is, indeed, underway and many industries have disappeared from remote parts of former bantustans'. Furthermore, Newcastle, for instance, is a sizeable town, and indeed is perhaps even regarded as a 'secondary city'. However, the fact remains that there is a huge discrepancy between the evidence, for Ezakheni and Newcastle, relied upon in the formulation of the current RIDP and the picture produced by the recent fieldwork in those centres.4 It suggests that some of the major premises underlying the current RIDP, described above, may have rested on shaky foundations. The evidence available to date, thus, does not seem to refute the view that there was a spontaneous tendency in the 1970s and 1980s for certain segments of manufacturing industry to shift towards labour abundant, low-wage areas. Indeed, lending support to this view is the fact that it has not been a peculiarly South African phenomenon. Hansen et al (1990:280), in considering international experience in the 1970s, states: 'Though it was not evident at the time, with hindsight, it appears that some spontaneous processes were already moving in directions consistent with certain of these [regional] objectives; in the 1970s other factors radically altered the regional environment'; and refers to 'favourable spontaneous ... changes ... in the locational preferences of people ... and of firms (manufacturing decentralisation) ...' Hart (1995:2) observes too: 'Since the 1970s and 1980s, tendencies towards industrial dispersal emerged in many different parts of the world - most dramatically, of course, in China where the massive surge in industrial growth has come largely from rural areas'. Also, consistent with this view is the fact that a significant part of industrial growth in some development points in the 1980s was the establishment of plants by East Asian investors. Platzky (1995:162-3) states: 'From 1982 industrialists from the Republic of China or Taiwan have accounted for some 60 percent of plants and jobs created in IDPs and about a third of the capital invested ...'. She explains this in terms of 'a major alliance forged in the 1980s' between South Africa and Taiwan, one element of which was that: Taiwan needed new sources of raw materials and markets for products not subject to import quotas, as well as cheap unskilled labour for relocated production in order to enter new markets ... By 1981 Taiwanese industrialists were considering establishing branch plants in RSA to manufacture knitwear, footwear, plastic and fabricated metal household goods. Since 1985 Taiwanese industrialists have relocated their operations to the bantustans in 12 TRANSFORMATION 32 (1997) BELL ARTICLE significant numbers ... In the late-1980s structural problems in Taiwan's economy forced industrialists abroad... Hence the move to open factories in Sri Lanka with eight-year tax holidays, where monthly wages [were] US$45 ... Thailand, the Philippines and Malaysia were also popular destinations for the export of Taiwanese capital as wages were low and their export quotas for the USA were still open..." (1995:162-3). This description of the involvement of Taiwanese firms in the industrial development points, provides one significant illustration that market forces were indeed at work, making for industrial growth in some development points in the 1980s. Indeed, it suggests one of several ways in which the emphasis on 'international competition', in the original argument for a spontaneous tendency towards industrial dispersal in South Africa in the 1970s and 1980s, was probably justified. To that significant extent at least it seems that (as the Panel of Experts recommended should be the case) regional industrial development policy in the 1980s was working with rather than against market forces. There is no reason to think that this did not also apply to a significant proportion of South African-owned plants which established or expanded in development points in the 1980s. Regional Policy as a Means of Promoting Structural Adjustment As noted earlier, two economic objectives of regional policy elsewhere are promoting structural adjustment and realising development potential more fully. On the first of these two objectives, Hansen et al (1990:282) state: 'Regional policy can be used to upgrade [low productivity, inefficient, traditional] industries'. Such a possibility was not considered in evaluating the 1982-91 RIDP, and in formulating the present one. As Hansen et al (1990:284) argues, however, 'regional policies need to be viewed in the light of changing industrial structures, as well as the changing determinants of location', and in this connection refers to 'the increasing spatial decomposition of production, the rise of "footloose" industries ... and the increasing importance of small and medium-sized enterprises ...'3 As the discussion above implies, given 'changing industrial structures, as well as the changing determinants of location' in South Africa, regional industrial development policy in the 1970s and 1980s (whatever its primary aims) may well up to a point at least have served the economically-useful purpose of facilitating structural adjustment in certain segments of manufacturing industry. To that extent it did not simply create distortions, resulting in a misallocation of resources, in head-on conflict with an improvement in the performance of TRANSFORMATION 32 (1967) 13 ARTICLE BELL manufacturing industry and the economy as a whole. In formulating the 1991 RIDP it seems that no consideration was given to the possibility of such a role for regional industrial development policy either in the 1980s or indeed in the future. What is the scope for regional industrial development policy playing such a role today? The strength of the attractive pull of low-wage sites for segments of manufacturing industry may well have lessened in the 1990s, for two reasons. One of these reasons is that adjustments to market forces, which, it has been argued, provided a large part of the initial impulse for dispersal to low-wage areas in the 1970s and 1980s, may at first be large, as they were especially in the 1980s, but tend to peter out with the passage of time. However, Hart (1995a:4), for instance, reports as follows on developments at Newcastle recently in response to new opportunities for clothing exports: ... China and other Asian countries confront quotas on their clothing exports to the United States. Goods produced in South Africa are not subject to quotas. Accordingly, even though costs of production in South Africa are higher than in most of Asia, the lifting of sanctions has made South Africa quite an attractive location. On the face of it, at least, Newcastle seems like a logical place for aspiring export producers to locate. Wages are substantially lower than in the major urban areas, and there are good road and rail links to the port in Durban and the international airport in Johannesburg. These are precisely the reasons why a large Singaporean clothing producer established in Madadeni in the first half of 1995, and has plans for significant expansion. The Newcastle local government and KwaZulu Marketing Initiative are negotiating with similar prospective firms and, like many others in the province, see this as the wave of the future. The market forces making for a tendency for plants to establish and expand at centres in and near the former homelands, thus, are evidently by no means completely spent. Regional industrial development policy may still have a role in this process. A second reason that the attractive pull of low-wage areas for certain segments of manufacturing industry may be weaker than before, has to do with intensified competition from imports. This is partly related to the lifting of sanctions, and partly to import liberalisation, through comprehensive tariff reductions, which is currently under way. While this will tend to reduce the effectiveness of regional incentives of any particular size, it would seem to provide an additional justification for the use of regional industrial development policy as a means of 14 TRANSFORMATION 32 (1997) BELL ARTICLE promoting an economically desirable process of industrial restructuring. It is to this that we now turn. Regional Industrial Development Policy in the Context of Changing Trade Relations Hart's very useful study provides an illustration of yet another phenomenon pertinent to the future of regional industrial development policy. She notes that the clothing industry in Newcastle is at present experiencing difficulties, and pinpoints two reasons for this. One is intense labour conflict. The second, relevant in the present context, is intensified competition from imports. She finds: 'Cheap imports, mainly from China, are stripping the profits of a large segment of the clothing industry in Newcastle - most notably knitwear - that has been the main source of growth over the past five years. This phenomenon is very recent, but is likely to gather force in the future. Many industrialists are threatening to close down and either move to other countries or switch to importing' (1995:4). The problem for manufacturers in Newcastle, due to the flooding of the domestic market with cheap imports, is so far probably mainly related to the lifting of economic sanctions, rather than to the liberalisation of South Africa's imports. The lifting of sanctions has made it possible for some foreign producers, many with cheaper labour than ours, but previously legally prohibited from trading with us, to sell their goods in the South African market for the first time. Whereas, as noted above, the lifting of sanctions may be having a beneficial effect on production for export in development points, such as Newcastle, its effect on production for the domestic market is negative. This problem is likely to be aggravated by trade policy reforms, which will result in a substantial reduction in levels of tariff protection over the next five to e ight years (See Bell and Cattaneo, 1996). Unskilled labour-intensive industries, such as clothing and textiles, are amongst the manufacturing sectors likely to be most severely affected by this.6 Clothing and textiles, though contributing a small proportion of manufacturing valued-added, employ a significant proportion (about 15 percent in 1993) of the manufacturing labour force. The effects of the tariff cuts on these industries thus must be a matter of great concern for the country as a whole. Some take a sanguine view of the likely effects of import liberalisation on s °uth Africa's unskilled labour-intensive industries. Kaplinsky (1994:535), elaborating on the position of the Industrial Strategy Project (ISP) in relation to Sl >ch industries, states: ' ... given South Africa's relative wage levels it is not possible to compete in a range of low- productivity, unskilled labour-intensive "TRANSFORMATION 32 (1907) 15 ARTICLE BELL industries without high levels of protection which undermine the consumption power of working-class incomes, as well as those of the unemployed ...'. Though no attempt will be made to establish this here, this argument for reduced levels of tariff protection for unskilled labour-intensive industries is completely inadequate. It is questionable in particular whether the tariff cuts taking place in these industries are in the interests of working-class people, including the unemployed.7 Furthermore, it might be noted that whereas the report of the ISP (Joffe et al, 1995:29) compares average wages in South Africa with those in the same sectors in other countries, wage rates at centres in and near the homelands are considerably lower than the South African averages in unskilled labour-intensive industries. Regional incentives which simply facilitate relative shifts to low-wage areas, by enabling plants to achieve labour cost savings, may in this way alone help some segments of these industries to increase their price-competitiveness. The scope for such labour cost savings alone to help maintain production capacity and employment in such industry segments, however, is being reduced by import liberalisation. As we have seen, even producers in low-wage areas are being adversely affected by intensified competition from imports. Taking the cuts in levels of tariff protection scheduled for 1995-2002 as given, thus, there would seem to be a case for providing additional protection in the form of regional incentives as a means of compensating, partially if not wholly, for the loss of tariff protection by segments of certain industries. Regional incentives, it seems, could serve as a GATT-compatible substitute for tariff protection, which could mitigate the adverse impact of import liberalisation on some vulnerable industries. This is economically desirable from the viewpoint of the country as a whole. It would also at least help preserve, if not increase, the level of manufacturing activity in some of the poorer regions of South Africa.8 From a social point of view, too, this is important. As noted earlier, the poorer regions of South Africa, including development points in and near the former homelands, are relatively specialised in unskilled labour-intensive industries such as textiles and clothing. So far as output and employment effects are concerned, thus, there is the danger that the incidence of the impact of import liberalisation, which will be especially severe in these industries, will be highly regressive (Bell and Cattaneo, 1996). Regional incentives could also help compensate in some measure for the loss of export subsidies provided in terms of GEIS, which is due to be phased out by the end of 1997. The Reynders Commission (1972) in fact proposed using decentralisation policy as a means of promoting exports, arguing that the 16 TRANSFORMATION 32 (1997) BELL ARTICLE government should provide additional incentives to potential exporters who moved to border areas. In those days, however, there was no restriction on the use by government of direct export-promotion measures, and it was therefore argued against Reynders that, in so far as the aim of policy was to promote exports, all exporters, wherever they were, should be subsidised on the same basis as regards their export activities (Bell, 1973). Indeed, until quite recently this seemed to be the proper approach. However, GATT has now largely removed that option. Though very much a second or third-best option, so far as export promotion is concerned, regional incentives may serve to maintain or increase the export potential of assisted regions. 6. Does a Regional Industrial Development Policy Targeting Low-wage Areas Really Benefit People in those Areas? One part of the case for the current RIDP was that, in view of the large inter-regional wage differentials, a job in an industrial development point in and near the homelands, was in some sense worth less than a job in a larger urban centre. This was apparently seen as a reason for not targeting such development points. It was recognised (POE, 1989:74) that one result of low wages was that the 1982-91 RIDP had relatively strong employment effects, and that together with the incentives, it provided the motivation for plants to locate in those areas. This gave people lacking in skills and mobility an opportunity to earn an income. Nevertheless, it was argued that the 'overall income effect is less than in major metropolitan areas and secondary cities' (POE, 1989:74); and that an 'implication of the existence of low wage levels at development points is that they are not providing a real alternative to metropolitan jobs or to migration ... men continue to migrate to metropolitan areas in search of jobs' (POE, 1989:74). That lower wage rates resulted in a lower (money) income effect than if the same job had been provided in a metropolitan area or secondary city is obvious. One implicit assumption underlying this view of the value of jobs in development points, however, seems to be that jobs created in terms of an RIDP are necessarily at the expense of employment in higher-wage metropolitan areas and secondary cities. This is completely unwarranted. One illustration of this is the fact, noted above, that a significant proportion of the additional output and employment provided in the low-wage development points was in Taiwanese plants relocated from outside the country, in order to take advantage of the cheap labour and subsidies available in the development points. They would almost certainly not otherwise have been established elsewhere in South Africa. TRANSFORMATION 32 (1997) 17 ARTICLE BELL Furthermore, though the wage rate may have been low, a job in an industrial development point clearly could have provided a very real, indeed most welcome, alternative to migration. This was obviously especially so where jobs were not available in higher-wage metropolitan areas and secondary cities. This was probably the case for most of the people employed in the development points in the economically-troubled 1980s. Even in the case where an individual had the option of a job in a higher-wage centre, however, employment in a low-wage area might have been preferred to migration, for various reasons - such as proximity to family, the costs of moving, maintenance of other sources of income in the home area, and crime in the cities. The fact that there was not sufficient growth in jobs in development points to make migration unnecessary altogether, does not detract from the value of jobs there for those who were able to get them. Platzky (1995), for instance, finds that thejobs created in terms of the 1982-91 RIDP were very beneficial, particularly to women, despite the low wages. She notes (1995:257): 'Until the establishment of IDPs the only cash income available to landless rural women living in either bantustans or the "white" farming areas was low-paid farm and domestic work'. She adds (1995:258): 'However minimal the cash wages, women in IDPs were able to earn some independent income and to exert more control over their households than they could as long distance migrants'; and Jaffee (1988:4) quotes an ex-Johannesburg domestic worker employed at Brits IDP who states: ' ... when I did domestic work, I had to leave my village in Bophuthatswana and I saw my kids monthly. Now that I work in the factories I return every night'. Ardington (1984:106), also quoted by Platzky (1995:257), had found somewhat earlier: The availability of local employment opportunities at Isithebe has clearly benefitted the entire community there, but the greatest improvement has been realised by women. There are virtually no local wage employment opportunities for the uneducated in rural areas and social norms and practical considerations often make it impossible for women to become migrants. In Isithebe this is no longer the case and women have not hesitated to become involved in the wage economy. Finally, though her focus is mainly on the benefits for women, Platzky (1995: 258) reports that 'male workers interviewed in Isithebe stated that the RIDP had been beneficial to them and had also prevented them having to seek work in the metro areas far from their homes'. In the formulation of the present RIDP, a consideration which apparently also influenced the abandonment of preferential incentives for the earlier development points was that they had largely involved the development of 18 TRANSFORMATION 32 (1997) ARTICLE BELL 'footloose' industries, which it was said have 'weak local linkages and hence poor propulsive characteristics' (POE, 1989:119). As noted earlier, however, the dispersal of such industries was in keeping with a worldwide tendency, both at the international level and, within countries, inter-regionally. Many of today's Newly Industrialising Countries began their phases of rapid industrialisation on the basis of unskilled labour-intensive, footloose industries. The initial impetus for manufacturing growth in the industrial development points - South Africa's newly industrialising areas - has also largely been provided by such industries. As in the case of the NICs, there could in future be a gradual upwards movement in the skill and technology hierarchy of manufacturing activities. As Hansen et al (1990:286) state: '... it is possible for a country or region to begin the development process by initially taking activities in the mature phase of the product cycle, but then to shift gradually the nature of its product and process mixes in favour of more sophisticated activities with greater skill requirements and higher wages'. 7. Agriculture versus Industry In arguing for the current REDP, it was stated (POE, 1989:220): 'Some of the nine development regions have potential for industrial development while others lend themselves better to development in the fields of agriculture, mining, forestry, tourism, small business and services'. No unambiguous indication was given of which regions and centres had potential for industrial development, but, as noted in section 3 above, such potential was seen as lying almost entirely in the country's secondary cities and metropolitan areas. Other places were regarded by and large as devoid of any significant comparative advantage in manufacturing industry. The other side of this coin was that places outside the secondary cities and metropolitan areas had a comparative advantage in one of the other sectors listed above. Agriculture was apparently seen as the major alternative to manufacturing. The present RIDP was not in fact grounded in any proper analysis of this issue. It was simply asserted in effect that the areas surrounding the development points of the 1982-91 RIDP in general had a comparative advantage in agriculture. It was stated (POE, 1989:131) that particularly 'over the past decade, increased attention has been focussed on broad-based assistance to small farmers as an instrument of agricultural, rural, and overall economic development'; and future development potential was evidently seen as lying mainly in that direction. At the time of the formulation of the 1991 RIDP there was considerable optimism about the potential for development of small-scale, black agriculture. This is reflected also in more recent studies such as by the World Bank (1993). TRANSFORMATION 32 (1997) 19 ARTICLE BELL This study contended that a land reform involving the subsidised transfer of 30 percent of the country's currently white-held commercial farmlands, would generate some 1.5 million new 'rural livelihoods' (a rural livelihood being defined by the World Bank as a certain amount of income, considered to be a full-time farming equivalent, which would vary from region to region). This it said could be accomplished in five years at a cost to the state of R17.5 billion. More sobering accounts of this potential, however, are to be found in more recent work by Hertz (1994), Murray (1995), Mclntosh and Vaughan (1995), and Hart (1995). Hertz (1995), though sympathetic to land reforms for the development of small-scale black agriculture, examines the World Bank's methods of calculation, and revises downwards the number of rural livelihoods that could be created by transferring 30 percent of white commercial farm land, from 1.5 million to 626 000. The total cost to the State for this smaller number of rural livelihoods is estimated by Hertz at R20.2 billion compared to the World Bank's estimate of R17.5 billion, and the cost per rural livelihood at R32,293 compared to the World Bank's estimate of Rl 1,846. It appears that both these estimates are based on the assumption that the productivity of small-scale agriculture will be as great as on white-held land at present, and that the terms of trade of agriculture will be unchanged. As Hertz notes, however, the substantial reductions in levels of price support, including reduced levels of protection against competition from imports, which are taking place, mean 'that black fanners will face prices that are lower than those enjoyed by the whites who came before them, and lower than those used by the [World] Bank in its forecasts' (1994:10). The cost to the State of creating a rural livelihood, thus, would it seems be greater even than Hertz's high estimate.9 Hart (1995:1) questions the case for a rural and agricultural strategy which involves 'creating peasants or small farmers wherever land becomes available through the market'. She argues that rural people with limited landholdings generally adopt 'multiple, diversified livelihood strategies', specialising completely neither in farming nor in wage employment, but deriving income from a number of different sources. An implication of her analysis seems to be that small-scale agricultural development alone is not sufficient to support rural households but has to be supplemented from other sources. Multiple livelihood strategies have long been prevalent in South Africa, but are not peculiar to this country. A difficulty with small-scale farming developments is that they are likely to be spatially dispersed, and to lack the opportunities for other livelihoods unless in close proximity to other forms of economic activity. This, Hart says, is why the small-scale farmer model is unlikely to be a realistic option for large 20 TRANSFORMATION 32 (1997) BELL ARTICLE numbers of poor people - particularly women - who even if they could get access to land would continue to rely on multiple livelihoods. Hart's argument also seems to imply, given the complementarity of different livelihoods, that where people in industrial development points do not have landholdings, the feasibility of providing them with small land units should be explored. Cultivation of these could provide some income, but people also require land as a form of social security. Indeed, she seems to say that the fact that people, particularly women, in Newcastle do not have land is one of the main reasons the Taiwanese model of development does not work in Newcastle (a centre with numerous Taiwanese-owned plants), and, in particular, why these plants have experienced severe labour conflict. Though she does not deal explicitly with this case, Hart's argument would also seem to suggest another argument for the promotion of industrial development in centres near existing small-scale, black farming areas. That is, where people already have land, industry adds another, complementary source of income. Mclntosh and Vaughan also express doubts on various grounds about the potential for small-scale agriculture, concluding that 'significant non-sustainable public investment will be required to generate a large small-holder sector outside those limited contexts where there is scope for such development' (1995:28). The present RIDP's assumption that the scope for small-scale agricultural development is large, was apparently influenced partly by the success of the small cane-grower sector in KwaZulu-Natal (POE, 1989:131). Mclntosh and Vaughan, however, regard this as an exception, as one of those limited contexts in which there has been scope for small-holder agricultural development. Also emerging from their discussion is the likelihood that market-based land reforms will result mainly in the transfer of white-owned commercial farms to large-scale black commercial farmers, rather than in the development of small-scale agriculture. One implication of this would seem to be that the transfer of 30 percent of currently white-owned land is likely to result in a much smaller, if any, increase in the number of net rural livelihoods than estimated by Hertz, with a corresponding further increase in the cost to the state per additional rural livelihood. At a rather broader level, the density of population in the rural areas of the formerly independent and self-governing homelands, alone, suggests strongly that those territories are unlikely to have a comparative advantage in agriculture. However, this question and the relative emphasis that should be given to agricultural and industrial development in South African regional policy need to be considered thoroughly. TRANSFORMATION 32 (1997) 21 ARTICLE BELL 8. The Swing of the Pendulum As Hansen et al (1990:5) state, 'regional policy reflects both what is happening within a national society and a national economy at any particular time and the prevailing economic and social philosophy at that time; and these interact with one another'. The introduction of both the 1982 and 1991 RIDPs, and the new direction which regional industrial policy is about to take in late-1996, can probably be understood only in such terms. The decision to adopt the policy which prevailed from 1982 to 1991 was taken in November 1981, at the 'Good Hope Conference', at almost the exact moment at which the gold-led boom of 1979-81 reached its peak, and before the economy had begun its rapid decline. This natural-resource boom resulted in a massive foreign exchange and fiscal windfall. As in the case of several other countries, particularly (but not only) oil exporters, which experienced such temporary natural-resource booms at about this time, South Africa saw these windfalls as an indication of a permanent favourable change in its economic fortunes. The fiscal resources available to the state seemed almost unlimited. The various countries which were beneficiaries of such temporary natural resource booms chose to use their windfalls in different ways, some wiser than others. The South African government chose inter alia to adopt a new RIDP with extraordinarily generous incentives to industrialists establishing or expanding plants in or near the 'homelands'. With the fiscal windfall, the bantustan system together with reform of the parliamentary system must have seemed after all to be a viable political option. By the mid-1980s, with the decline in the gold price, and other untoward economic events, as well as widespread unrest in urban black townships, it was clear that the growing cost of the RIDP was becoming both unaffordable and politically pointless. The process of extricating the country from the 1982 RIDP, however, took a considerable time. In 1986 the critical evaluation of the RIDP by the 'Panel of Experts', managed by the Development Bank of Southern Africa began, and its report was submitted to the government in January 1989 (Platzky, 1994:121). Given the turn of economic and political events, the 1982 RIDP was clearly unsustainable. As Platzky (1994:121) notes, by the end of the 1980s, the RIDP was costing the taxpayer Rl billion a year, and relative to this expenditure was generating few jobs in the industrial development points. The system of incentives was open to abuse and there was a growing conviction that corruption was costing the taxpayer millions more than necessary (Platzky, 1994:125). There was probably also some validity to criticism of the number and location of some of the industrial development points (Platzky, 1994:123). 22 TRANSFORMATION 32 (1997) BELL ARTICLE South Africa's regional industrial development policy was therefore clearly in need of change. The above critique of the thinking underlying the 1991 RIDP does not deny this. It is held only that the attack of the Panel of Experts on some of the basic principles underlying the 1982 RIDP, on the grounds that they were neither compatible with economic efficiency, nor necessary on social grounds, was logically flawed. Amongst these principles, the most fundamental was that of preferential treatment for economically disadvantaged centres in or near what are now the former homelands. A consequence of this attack on fundamental principles was that, instead of reform of the 1982 RIDP, which was badly needed, the baby was in effect thrown out with the (admittedly murky) bathwater. Failure to consider seriously the possibilities of reform and thus to retain some of the basic principles of the 1982 RIDP, appropriate to a country like South Africa, was probably an over-reaction to the excesses of the policy, and its historical association with apartheid, but was also due to the growing receptiveness to arguments, however weak, which expressed unqualified faith in the efficacy of market forces. As Platzky (1994:121) notes, too, by the mid-1980s both capital and labour, for different reasons, found themselves in opposition to the RIDP. These forces are reflected also in the change in the nature of the regional incentives in 1991, which were related mainly to profitability, and involved a considerable de-emphasis of employment creation, a major focus of the 1982 RIDP, as a determinant of the level of financial assistance. In the process, however, it in effect represented a withdrawal from regional policy, as traditionally conceived and practiced round the world. Since the 1960s regional policy worldwide has evidently been subject to periodic swings, and this applies also to South Africa. Hansen et al (1990:281) say of regional policy in developed countries in the 1970s: "The activism of the 1960s was replaced by new faith in the efficacy of market forces; the pendulum had swung once again from interventionism to non-interventionism as the guiding philosophy'. However, he adds, the 'story in the LDCs is different because, although they too suffered from the global recession, they did not abandon national development planning, nor the regional component of it. Indeed, in many of them regional planning increased in the 1970s'. This is true also of South Africa. Efforts at regional industrial development were increased both in the 1970s and 1980s. The pendulum, however, did eventually swing in a non-interventionist direction in the case of South Africa, and our policies were modified accordingly in 1991. According to Hansen (1990:287-90), 'the general socio-economic and political climate may well be due for a new swing in the direction of greater activism ... There are various forces at play leading TRANSFORMATION 32 (1997) 23 ARTICLE BELL governments into regional policies and programmes ... government will intervene ... and will 'not stand idly by and allow the unfettered market to call the adjustment tune'. The pendulum is evidently also swinging back in South Africa at least part of the way, so far as some basic principles are concerned. Just as the policy shift of 1991 cannot be understood simply in terms of the persuasiveness of the arguments advanced for it at the time, however, so (regrettable as this may be) the shift in thinking now taking place cannot be attributed to superior scientific insight pure and simple. In accordance with the observation of Hansen et al, it reflects the socio-economic and political climate prevailing in South Africa today. 9. Implications for a Future Regional Industrial Development Policy As noted at the outset, regional policy elsewhere in the world, as in South Africa before 1991, generally targets particular regions. As Bloch (1993:23) observes, the 1991 RIDP represents a 'scattershot approach ... that... does not target either specific industrial sectors, or particular places ...'; whose 'distinctiveness as regional industrial policy [is] tenuous' (1993:24, footnote 28); and which is 'ironically "aspatial"'. In targeting particular regions, regional policy generally seeks to contribute to a reduction in inter-regional income inequalities, as well as in the unemployment rate and the incidence of poverty, in areas where these are most severe. The 1991 RIDP like its predecessors was not directed towards promoting such social objectives. This, it seems, is about to change. Inequality, unemployment and poverty are clearly central to the socio-economic philosophy of the present South African government, as expressed for instance in the Reconstruction and Development Programme (RDP). This is bound to affect regional policy. As Platzky (1994:128) states too, 'regional development is politically necessary in South Africa in order to provide for the development of those who have historically been excluded from privilege'. While homeland governments have gone, the existence under the present constitution of nine fully-fledged regional governments, with considerable powers, has probably reinforced the need for the central government, for political reasons, to pay a good deal of attention to the geographical distribution of economic activity. As also noted at the outset, regional industrial development policy may also be aimed at promoting structural adjustment, and realising a country's potential for economic development more fully. Though this was clearly not their main 24 TRANSFORMATION 32 (1997) ARTICLE intent, previous South African RIDPs may well have had the effect of facilitating an economically necessary process of adjustment in certain important sectors of manufacturing industry. Earlier regional industrial development policy may, thus, have served as a useful instrument of industrial policy. Over the past few years, however, there has been an increasing emphasis in South Africa on the need for 'industrial restructuring', and this too it seems is having an influence in the re-shaping of regional industrial development policy. What the scope is for regional industrial development policy for this purpose it does not seem possible to say in advance. In low-wage areas, as noted above, it may for various reasons be smaller than it was in the past. Adjustments to market forces, which it has been argued provided a large part of the initial impulse for dispersal to such areas in the 1970s and 1980s, may at first be large, as they especially were in the 1980s, but tend to peter out with the passage of time. Furthermore, as noted earlier, there are now two significant, relatively new factors: import liberalisation through substantial declines in levels of tariff protection, and the lifting of trade sanctions. Regional industrial development incentives of any particular size, thus, are likely to produce a less significant response, in terms of industrial expansion in low-wage areas, than they would have done in the 1980s, because of intensified competition from imports. This is particularly so since the magnitude of the tariff cuts due to take place between 1995 and 2002 tends to be greatest in unskilled labour-intensive industries, in which, as noted above, low-wage regions are relatively specialised. To counter the likely adverse effects of import liberalisation on import-competing manufacturing industries, the South African authorities would seem so far to have wanted to rely largely on supply-side industrial policy measures, such as those proposed by the ISP. However, there are great uncertainties about the likely effectiveness of these supply-side industrial policy measures, as a means of increasing the international competitiveness of South African industry. No serious consideration has been given to the question whether the supply-side measures can, not only raise the efficiency of manufacturing industry sufficiently, but do so quickly enough to stave off the ill-effects on certain industries of intensified competition from imports. Regional incentives thus may have a useful role as an alternative or additional means of facilitating structural adjustment in certain industries. Given the Government's commitment to the use of supply-side industrial policies, it is hard to see any objection in principle to the use of regional incentives for this purpose. Both regional incentives and supply-side measures are interventionist. Both involve the use of scarce fiscal resources. There is nothing TRANSFORMATION 32 (1997) 25 ARTICLE BELL to choose between them on these grounds. The choice is a purely pragmatic one: cost-effectiveness. In the context of trade policy reform, both supply-side industrial policies and regional incentives may be seen as substitutes for tariff protection (and indeed also GEIS). With reduced freedom to use traditional trade policy measures, regional incentives provide a GATT-compatible means of giving continued protection for certain manufacturing sectors. However, today the country faces a much tighter fiscal constraint than it did when the previous RIDP was introduced in 1982. As noted earlier, in the wake of the 1979-80 gold-led boom, the resources available for regional industrial development seemed almost unlimited. The result was an extraordinarily generous system of incentives which was bound to be unsustainable. Today, though there was an improvement in the economy in 1993-95, this came after several years of economic decline, and there are now also many additional demands on the State's finances. The outlook for the economy has also worsened considerably since February 1996. The level of incentives available to firms in and near the former homelands in a future RIDP, thus, will probably have to be relatively modest. However, the earlier incentives were doubtless unnecessarily generous, and almost certainly fiscally wasteful, even if they did not result in a substantial real misallocation of resources in manufacturing activity. More modest, well-structured incentives, could probably still have a significant effect. Fewer, and more careful selection of places eligible for regional incentives would also help contain the fiscal cost, and increase the efficacy of the incentives; though, under the new constitution, limiting the number of development points may prove politically difficult. Some suggest that direct transfers from central government to the provinces would be more efficient than regional industrial development policy as a means of dealing with our regional problem. For instance, Maasdorp (1990), cited by Bloch (1993:76), proposes 'direct income transfers ... to lagging regions ... to equalise standards of education, health, transport etc. between regions'. Such equalised transfers across regions, however, do not deal with inter-regional inequalities, which are due to differences in access to jobs and income.10 None of the above discussion implies that industrial development should be the only instrument of regional development policy. Nor is it implied that industrial development on its own can solve the problems of South Africa's poorer regions, or perhaps even make a major difference to conditions in these regions. However, neither this nor vague references to multi-sector development is a justification for doing nothing so far as regional industrial development is concerned. Regional industrial development policy, though inadequate on its 26 TRANSFORMATION 32 (1997) ART1CLE BELL own, should be taken as far as can be justified, giving due consideration to the costs involved, on economic and social grounds. 10. A Postscript As part of the Government's Manufacturing Development Programme (MDP), legislation has been passed which provides for 'tax holidays' for new industrial enterprises with assets in excess of R3 million. This provision has in common with other parts of the MDP the aims of raising fixed investment in manufacturing, restructuring domestic manufacturing towards international competitiveness, and facilitating a higher degree of international competitiveness. Its primary aim thus is an improvement in the output and employment growth performance of South African manufacturing industry in general. The provision for tax holidays, however, has a spatial component. A new manufacturing plant with assets over R3 million qualifies for a tax holiday of two years each for satisfying each of the following criteria: location in one of a number of designated places or areas; belonging to one of a number of designated industrial sectors; and a labour cost to value-added ratio exceeding 55 percent. A new manufacturing plant thus can in principle qualify for a tax holiday of up to six years. South Africa's new system of regional incentives thus consists essentially of this spatial component of the tax holiday. Section 37(H) of the Income Tax Act, which provides for the tax holiday, states that in designating qualifying locations, regard will be given to the following: (i) reinforcement of secondary cities; (ii) reinforcement of key urban nodes along any development corridor which qualifies as a spatial development initiative; (iii) consolidation of emerging agglomeration areas where a sufficient supply of appropriate infrastructure is available; and (iv) supporting of the diversification of local economies where a sufficient supply of appropriate infrastructure is available as a result of the restructuring of existing manufacturing activities. Though the list of places qualifying for the spatial component of the tax holiday has at the time of writing not been published, it appears that a substantial number of locations (over 50) is under consideration. On what basis each place or area on the provisional list has been chosen, and in particular what relative weights have been attached to social and economic criteria in selecting them, is unclear. The two-year spatial component of the tax holiday seems unlikely on its own to have a significant impact on the ability of the designated, eligible centres to attract new plants. The likely overall effect seems to depend on the extent to TRANSFORMATION 32 (1997) 27 ARTICLE BELL which these designated places also have locational advantages in the designated industries, including in particular the extent to which production processes in such industries are sufficiently labour-intensive to give a labour cost/value-added ratio in excess of 55 percent in those places. That is, the significance of the tax holiday provisions for the location of industry in the designated places, depends on the proportion of new manufacturing plants in the country as a whole for which these places are likely to qualify for a tax holiday, of not merely two years, but of four or six years. In the present context, however, noteworthy features of the new approach, by contrast with the 1991 policy, are targeting of particular places or areas, the renewed emphasis on employment creation, and the apparently deliberate attempt to use regional policy as an instrument of industrial restructuring. At least in respect of the first two of these features, the new approach seems to represent a shift back towards some of the basic principles of the 1982 RIDP, though without having anything like the same power to stimulate industrial development in the targeted places and areas." NOTES 1. This is a revised version of a paper written originally as a contribution to the recent evaluation of South African Regional Industrial Development Programme. The revisions have been made mainly in response to comments by an anonymous referee, and a conversation with Laurine Platzky, for which I am most grateful. 2. If the divergences associated with infant industries and infant areas arguments do vary systematically in the manner suggested, the present system of uniformly applicable subsidies would seem to be biased against development of the long-term industrial potential of centres in or near the former homelands, relative to secondary cities. The drafters of the present policy (POE, 1989:106) thus were probably right in seeing their proposals as effectively targeting 'the country's secondary cities, not small towns or isolated locations'. 3. As, for instance, in Bell (1983). 4. Furthermore, the 'major shaking out process' to which Hart (1995) refers may have been to a significant extent due to the intensified competition from imports (see Bell and Cattaneo, 1996), and the very unfavourable conditions for manufacturing industry in the country as a whole, in 1990-93, which resulted in a sharp decline in manufacturing employment. 5. Hansen et al are here apparently thinking mainly of developed countries and the 'footloose' industries referred to thus are those which 'depend on an innovative milieu'. The basic principle that regional policy should take into account the changing determinants of location and the other spontaneous trends to which they refer, however, applies generally. 6. For instance, the tariff rate on articles of apparel and clothing accessories fell from 106 percent in 1994 to 90 percent in 1995, and is due to fall further, to 40 percent, by 2002. 7. See Bell and Cattaneo (1996) for fuller discussion of this issue. 28 TRANSFORMATION 32 (1997) BELL ARTICLE 8. This clearly could not insulate vulnerable industries altogether from the adverse effects of import liberalisation. Most plants in major urban centres probably would not find it feasible to move. There could also be resistance from workers in these larger urban centres, and perhaps an attempt to thwart the process. 9. How this compares with the cost of generating income, equal to that of a rural livelihood, through the promotion of industrial jobs in and near the former homelands, is not known, but should be calculated to obtain a more definitive view on the case for industrial relative to agricultural development. 10. Curiously, some writers who seem to prefer direct income transfers to regional incentives aimed at social objectives within South Africa, nevertheless apparently accept such a regional policy on a sub-continental scale. For instance, Bloch (1993:76) says: If the country, like contemporary Europe, truly emerges as a 'South Africa of the Regions', then regional objectives akin to those funded through the European Community's Structural Funds (eg. the Economic Regional Development Fund, the European Social Fund) assisting the development of lagging regions; alleviating the difficulties of declining industrial regions; fighting long-term unemployment - might also be considered and capacitated. 11. Another significant aspect of South African regional development policy is the Spatial Development Initiatives (SDI) programme (Goldin and Jourdain, 1996). This is not completely unrelated to the spatial aspect of the MDP. As noted above, situation in a SDI is one criterion in selecting places which qualify for the spatial component of the tax holiday. However, apart from the fact that the SDI programme is multisectoral, rather than focused exclusively on manufacturing, the theoretical and policy issues which it raises are very different from those considered above, and do not fall within the scope of this paper. REFERENCES Ardington, E (1984) 'Decentralised industry, poverty and development in rural KwaZulu1, Carnegie Conference Paper No 246. Bell, T (1973) 'Some aspects of industrial decentralisation in South Africa", South African Journal of Economics, 41(4). Bell, T (1983) 'The growth and structure of manufacturing employment in Natal, Occasional Paper No 7, Durban: Institute for Social and Economic Research, University of Durban- Westville. Bell, T (1987) 'Is industrial decentralisation a thing of the past?', in R Tomlinson and M Addleson (eds). Regional Restructuring under Apartheid: urban and regional policies in contemporary South Africa, Johannesburg: Ravan Press. Bell, T and N Cattaneo (1996) 'Foreign trade and employment in South African manufacturing industry', Labour Market Policy Studies, International Labour Office, Geneva. Bloch, R (1993) From Dispersal to Concentration; Regional Industrial Development Policy in South Africa: past, present andfuture. Industrial Strategy Project, University of Cape Town. Centre for Development and Enterprise (1995) Post-Apartheid Population and Income: a new analysis, CDE Research No 1, written by J McCarthy and A Bernstein. TRANSFORMATION 32 (1997) 29 ARTICLE BELL Goldin, I and P Jourdain (1996), 'Development corridors: their contribution to investment and employment creation', Paper presented at the Southern Africa Trade and Investment Summit, Harare, 23 October. Hansen, N, B Higgins and DJ Savoie (1990) Regional Policy in a Changing World, New York and London: Plenum Press. Hart, G (1995) 'Beyond the rural-urban economy: rethinking agrarian reform in South Africa', Paper prepared for the Congress on Agrarian Questions, Wageningen Agricultural University, The Netherlands, May 22-24. Hart, G and A Todes (1997) 'Industrial decentralisation revisited', Transformation, 32. Hertz, T (1994) '1.5 million jobs in five years?, Comments on the World Bank's November 1993 land reform proposal', Paper presented at the Economic History and Development Workshop, April 20, 1994, as revised August 17, 1994. Jaffee, G (1988) 'Commuter labour: changing women's lives, changing households', in Agenda, 3. Joffe, A, D Kaplan, R Kaplinsky and D Lewis (1995), Improving Manufacturing Performance in South Africa, Report of trie Industrial Strategy Project, University of Cape Town Press, for the Industrial Strategy Project and the International Development Research Centre. Kaplinsky, R (1994) 'Economic restructuring in South Africa: the debate continues: a response', Journal of Southern African Studies, 20(4). Maasdorp, G (1990) 'Introductory survey', South African Journal of Economics, 58(2), Special issue on Regional Policy in South Africa. Mclntosh, A and A Vaughan (1997) 'Enhancing rural livelihoods in South Africa: myths and realities'. Revised version of Paper presented at the Rural Livelihoods Workshop, Durban, December 1995, in Land, Labour and Livelihoods in Rural South Africa: Volume 2 KwaZulu-Natal and Northern Province, Michael Lipton, Merle Lipton and Frank Ellis (eds), Durban: Indicator Press. Murray, C (1995) 'Who are the 'farmers'? The allocation of state-owned land in South Africa: a case study of land redistribution in the Orange Free State', mimeo. Panel of Experts (1989) Report on the Evaluation of the Regional Industrial Development Programme as an Element of the Regional Development Policy in Southern Africa, Halfway House, Development Bank of Southern Africa Platzky, L (1995) 'The Development Impact of South Africa's Industrial Location Policies: An Unforeseen Legacy', Unpublished DPhil thesis, Institute of Social Studies, The Hague, The Netherlands. Reynders Commission (1972) Report of the Commission of Inquiry into the Export Trade of the Republic of South Africa, RP 69/1972 Pretoria: Government Printer. World Bank (1994) 'Options for Land Reform and Rural Restructuring in South Africa', Revised version of paper presented at the Land Redistribution Options Conference held in Johannesburg, October 12-15, 1993. 30 TRANSFORMATION 32 (1997)