DEBATE IMPROVING MANUFACTURING PERFORMANCE IN SOUTH AFRICA: A REJOINDER Trevor Bell As noted in my paper in Transformation 28 (Bell, 1995), the Report of the Industrial Strategy Project (Joffe et al, 1995) deals with questions important for the future of the South African economy, and its views were likely to have a significant influence on South African policy. For these reasons the Report warranted careful attention, and I was pleased that two of it's authors, David Kaplan and David Lewis, had responded to my assessment of the Report. Constructive debate on these issues is important. Regrettably, Kaplan and Lewis have not provided any new substantive arguments, directed at the issues raised in my paper, to deal with in this rejoinder. Rather, their remarks indicate that in several crucial respects they have missed the point of my argument and hence misinterpreted it. The discussion which follows is devoted to showing this, and to clarifying the issues raised by the ISP report. Only once this is done is it possible to debate further the issues central to the formulation of South African industrial policy. For ease of reference, the discussion below considers the issues raised by Kaplan and Lewis under their section headings. 'Disputing the Productivity Performance. Was it really that bad?' On this aspect of my paper, Kaplan and Lewis' response consists mainly of references to the findings of studies not mentioned in my paper, in particular Moll (1990), the National Productivity Institute (1990), and the World Bank study by Belli et al (1992). I was in fact well aware of these studies, which are also discussed in Wright (1993:2-3, 11), a study of which I made considerable use in my paper. I did not refer to these findings, however, because they do not seem to me to throw light on the issues raised by the ISP, which I was addressing. TRANSFORMATION 31 (1996) 43 DEBATE BELL For instance, as Kaplan and Lewis state (1996:116), two features of Moll's findings are a consistent declining trend in manufacturing TFP growth after 1954-63, and strong negative growth in the 1980s. Neither the declining trend nor negative TFP growth in the 1980s were at issue in my paper, though whether this negative growth could be described as 'strong' is more debatable.1 The only question was whether these features were peculiar to South Africa. In the context of the discussion in the ISP report (1995:3-10), it seemed these were seen as distinctive features of the TFP growth performance of South African manufacturing industry. That they are seen as distinctive is explicit in the case of the negative TFP growth rates in the 1980s, which are described (1995:11) as 'highly unusual'. The evidence provided in the ISP report, consisting solely of a table showing the economy-wide (as distinct from manufacturing) TFP growth rates of OECD countries (rather than of comparable developing countries), could not establish this. I therefore set out to consider in some detail the available relevant evidence, something the ISP should itself have done in the first place. It turned out that there was a general tendency for TFP growth rates to decline and that negative TFP growth rates were not highly unusual. This was all I claimed, and Moll's findings as reported by Kaplan and Lewis do not contradict this claim. Kaplan and Lewis (1996:117) state further that the studies they cite 'all concur that [manufacturing] productivity performance has been poor (very low positive or negative) over a long period and that [this] is a distinctive or striking feature which requires explanation'. This is apparently intended to support the claim in the ISP report (1995:11) that South African manufacturing industry 'has had a very low rate of productivity growth for a sustained period'. How low is 'low'? 'Low' is a relative notion. Something can be said to be low only in relation to some objective standard, in particular in the present context, in relation to other comparable countries. The claims of the ISP that the TFP growth of South African manufacturing has been 'low', and now of Kaplan and Lewis that it has been 'poor', either mean low or poor in this relative sense or they mean nothing. The ISP did not provide evidence on the TFP growth record of South African manufacturing relative to other countries. I therefore considered such evidence as was available to me, and found that it could not sustain the notion that the TFP growth performance of South African manufacturing industry was particularly low compared to developing countries in general. The studies now cited by Kaplan and Lewis (Moll 1990, NPI 1990, and Belli et al 1992), though they may 'all concur...', do not contradict this. Kaplan and Lewis (1996:117) complain: 'Bell never makes mention of the sectoral studies which underpin the ISP - but micro-sector and firm-level studies, 44 TRANSFORMATION 31 (1996) BELL DEBATE and not the macro, is the terrain that the ISP study essentially traverses'. It was however the ISP report itself that in the first place gave centre-stage to macro-level TFP growth. In the face of criticism of their claims on the basis of macro-level evidence they now shift their ground. While it is true I do not mention the ISP's micro-sector and firm-level studies, it seems that neither the ISP report nor Kaplan and Lewis do more than merely mention them. In the ISP report (1995:14), studies of sectors and firms are mentioned in one short paragraph, after a discussion of productivity which runs entirely in terms of TFP growth. Kaplan and Lewis (1996:117) state only that 'at the firm level, in most manufacturing sectors, the ISP found low levels of productivity and, even more important, an increasing tendency for most South African firms to fall further behind the leaders...'. Neither they nor the ISP report provide any elaboration, let alone attempt to demonstrate this. If they could show on the basis of their sectoral studies that the levels of productivity in South African manufacturing in general are lower, and have grown at a slower rate than in other comparable countries, this might strengthen their case. Given the well-known difficulties involved in cross-national micro-level productivity comparisons, it would also represent a significant scientific achievement. Why has this not been done - particularly considering that this essentially is their terrain? Nothing in the above argument implies that the TFP growth performance of South African manufacturing industry has been good relative to other comparable countries, or even that it was not 'really that bad' or 'not that weak' (1996:120). The ISP report, and Kaplan and Lewis, may well be right in claiming it has been poor relative to such countries. I contend only that neither the TFP growth evidence in the ISP report and in Kaplan and Lewis' response, nor the evidence which I considered in my paper, allow one to say this. This part of the discussion of the ISP and Kaplan and Lewis thus provides no support for their emphasis on slow productivity growth as the key to either poor output growth relative to other developing countries in the past or to an improvement in future. 'Explaining Productivity Performance. But does it really matter anyway?' In this section, Kaplan and Lewis (1996:118) state: 'Bell sees productivity growth as determined solely by macro-economic factors'. This is a grossly inaccurate representation of what I actually said. In criticising this aspect of the ISP report, I state (Bell, 1996:6): 'Virtually no account is taken of the impact of macro-economic factors on the TFP growth TRANSFORMATION 31 (1996) 45 DEBATE BELL rate'. Whereas for the ISP, in the statistical association between productivity growth and output growth the 'direction of causation runs mainly from productivity growth to output growth', I pointed out that to 'a large extent..., at least in the short to medium-term, the direction of causation runs the other way [i.e. from output growth to productivity growth]'. In support of this, I quote Helleiner (1994:29), who on the basis of detailed studies of fourteen developing countries states that 'the growth of output itself was found to be the 'principal concomitant of manufacturing productivity growth', and that: 'Much of mis association between productivity growth and output growth is attributable to the impact of variation in capacity utilisation'. I also quote Ocampo (1994:162) who says: "TFP is largely a dependent rather than an independent variable' (all emphasis in the above has been added). They do not say 'solely', and neither do I. Kaplan and Lewis' mis-statement of my position diverts attention from their neglect of this important consideration, by attributing to me an utterly indefensible, mono-causal view, in which the direction of causation runs entirely from output growth to productivity growth, and in which other determinants of productivity growth do not count at all. My fundamental criticism of this aspect of the ISP report was that it gives 'no systematic consideration ... to the determinants of manufacturing TFP growth'. This criticism stands. Kaplan and Lewis fail to appreciate its significance. The ISP report states (1995:22): 'The principal objective of our industrial strategy is to raise the productivity of South African manufacturing firms'. Given this, some reasonable notion of the determinants of TFP growth (whether at the macro or micro-level) is clearly essential. The basic principle is the unexceptionable one that diagnosis of ills must precede prescription. If it is true, as the evidence suggests, that to a large extent the TFP growth rate depends on macro-economic performance, at least in the short and medium-term, then it follows as I stated (Bell, 1996:7) that 'micro level efforts to increase efficiency may have little effect on productivity growth in the absence of a significant improvement in macro-economic conditions'. Also in this section, Kaplan and Lewis (1996:118) level a second, distinctly different criticism, concerning my view of the role of productivity growth in the longer-term. They claim, for instance, that for me even in the long run 'productivity growth is unimportant' as a determinant of output growth, and (1996:119) attribute to me the 'contention that the level of investment is all important'. This creates a misleading impression of my argument. On the basis of certain research findings published in the general literature, I stated that 'TFP growth as 46 TRANSFORMATION 31 (1996) BELL DEBATE such is not the major part of the explanation of why growth rates differ internationally'; and that 'productivity growth is not the only, or indeed, even the major, determinant of output growth performance, which depends heavily on a host of other factors'. Relying primarily and directly on Young (1994) (not, contrary to the impression given by Kaplan and Lewis, primarily on Krugman or the other works mentioned by them), I interpreted Young's work as showing that 'contrary to popular opinion ... what has distinguished the East Asian NICs from other developing countries, and enabled them to achieve higher per capita income growth than any other country in 1960-85 ... has not been exceptionally high TFP growth rates, but huge increases in capital and labour inputs'. Neither Young's findings, nor my discussion based on them, imply the views Kaplan and Lewis attribute to me. They mean only what they say, that factor input growth, that is, increases in both capital and labour, as Young shows, are a more important part of the explanation, in a growth accounting sense, of cross-national growth rate differences, and indeed are a more important part of the explanation of the superior growth performance of the East Asian NICs than differences in TFP growth rates. This does not necessarily mean that productivity growth is an unimportant factor in growth performance. It means only, as I have said, that, according to Young, it is less important than factor input growth in explaining why growth rates differ internationally, and in particular why the East Asian NICs have grown faster than virtually all other countries. If even this more accurate version of what Young and I say is unacceptable, Kaplan and Lewis should direct their attack at Young, of whose bad tidings I was simply the messenger.2 What is it about the findings reported on pages 5-8 of my assessment of the ISP report that makes Kaplan and Lewis resist them so strongly? Is it that the ISP has committed itself to a policy position which in their view does not allow of any significant determinants of output growth other than productivity growth? It seems that the arguments advanced above and in my assessment of the ISP report do not necessarily undermine the importance of productivity growth or of efforts to increase it. They do suggest, however, that there is a good deal more to raising productivity than the ISP's supply-side industrial policy increases; and that the attainment of substantially higher output-growth rates depend on an industrial policy which does more than simply aim to increase productivity. TRANSFORMATION 31 (1996) 47 DEBATE BELL 'Providing an Explanation for the Slower Growth of Manufacturing Industry. What accounts for the level of investment?' In this section, Kaplan and Lewis state that for me the 'key causal explanation ... lies with low levels of investment. Investment is all'. In fact, it is the ISP itself which asserts that low investment is one of the two outstanding factors (the other being low productivity growth) that 'have contributed to poor manufacturing performance'. Far from seeing it as a 'key causal explanation', I argued (Bell, 1996:10) that the 'decline in the I/GDP ratio in South Africa in the past ten years or so, rather than being a cause, is a symptom of the same macro-economic problems ... which produced the deterioration in the performance of manufacturing industry'. My discussion of investment deals solely with the causes of the decline in the I/GDP ratio, not with this ratio as a cause of anything. My basic criticism of the ISP report is that it completely neglects this question. The ISP report refers (1995:19) to 'political and macro-economic factors', but only in passing, and without elaboration. Their view of the causes is presumably put forward by Gelb (1991) to which they refer me. Gelb is simply an afterthought in the face of my criticism, to justify their failure to address the causes of the decline of the I/GDP ratio. It is indeed doubtful whether this demanding task really interests them for, after referring to Gelb, they give as their reason for neglecting this question that they 'saw no reason, in an industrial strategy project, to revisit this exercise' (emphasis added). This apparently is the basic problem. They have from the outset seen the ISP's business as that of producing policies aimed at raising productivity directly through industrial restructuring, by means of supply-side measures. Other aspects of the poor performance of South African manufacturing are seen as lying outside their terms of reference. Given this, there cannot be any rivals to micro-level industrial policies. If, notwithstanding this, Kaplan and Lewis are interested in this question, it seems not unreasonable to suggest that, instead of saying in effect that I should have read everything written by a member of the ISP (in particular Kaplinsky 1995), or of any of its antecedents (in particular Gelb 1991), they should produce a coherent criticism of my position (not just vague references to the 'social regime of accumulation' and apartheid), and a statement of an alternative view, based on these authors and any other relevant material. Then it will be possible to debate further the substantive issues related to the problem of the determinants of the level of investment in South Africa. 48 TRANSFORMATION 31 (1996) BELL DEBATE 'Exports and Imports. Grounds for optimism or pessimism?' In this section, Kaplan and Lewis state (1996:122): 'Successful performance on domestic markets cannot substitute for exports. We have to increase our penetration of international markets. How does Bell imagine that we are going to achieve the requisite rate of output growth ... if not through exports? ... the macro-economics of an inward-oriented industrial growth path are not sustainable'. They also refer to my 'more or less exclusive focus on the domestic market', and (1996:127) to 'the autarky proposed by Bell', and say (1996:124) that I contest 'the outward-orientation of the ISP'. The clear implication of this is that I have denied the necessity of accelerated export growth, that I advocate an inward-oriented industrial growth path, and that my emphasis is more or less exclusively on production for the domestic market, to the point of proposing autarky. This is a gross mis-statement of my position. I state (Bell, 1996:17): 'The issue ... is not whether South Africa should in future pursue a strategy of ISI or one of export-oriented industrialisation (EOI). As argued elsewhere (Bell, 1993), South Africa was forced by the macro-economic shocks of the early- to mid-1980s to shift abruptly from ISI to EOI, and there can be no going back on that'. In the 1993 study to which reference is made here, I stated (1993:99): 'South Africa has been forced to shift to an outward orientation primarily because of the need for accelerated export growth to overcome a foreign exchange constraint... In such conditions ... the question of the effect of protection on export expansion is ... crucial. Policies which prevent faster export growth could condemn South Africa ... to many further years of stagnation'. The question which I raise in my assessment of the ISP report, however, is whether, even when all policy obstacles to export expansion have been removed, the rate of export growth can be sufficient to allow South Africa to sustain some desired, higher GDP growth rate (of even 3.5 per cent, let alone the more ambitious growth rates now bandied about), given the economy's current degree of dependence on intermediate and capital goods imports, as measured by the import/GDP ratio. This is a perfectly meaningful question which poses an analytical and policy challenge. My claim in this regard (Bell, 1996:17) was that 'there is the distinct possibility that the required export growth rate, as defined above, will exceed the maximum feasible rate'. My criticism (1996:17) was that despite their 'implicitly rather un-reassuring view of the prospects for export expansion, the ISP does not consider this possibility'; and that the ISP 'in effect simply assumed that exports will be sufficient for a significant sustained increase in the rate of growth of manufacturing output'. Kaplan and Lewis could quite reasonably have tried, on TRANSFORMATION 31 (1996) 49 DEBATE BELL theoretical or empirical grounds, to deny that this is a real possibility, but they do not. Instead, they miss the point of the discussion, and thus fail to see the issue. The issue then, is this: What if export expansion alone (the 'more or less exclusive focus on exports' of Kaplan and Lewis), together with sustainable capital inflows, is not enough, given our current dependence on imports? Either we must give up our aspirations to a substantially higher GDP growth rate, or reduce the import/GDP ratio applicable at such a higher growth rate, through import substitution. There are no other alternatives. This does not mean that I contend that 'further ISI is possible', or that we should 'maintain our historic commitment to ISF. This should have been clear from my statement that 'there can be no going back' on our shift from ISI to EOI. That they nevertheless attribute these views to me appears to stem from their confusion of 'import substitution' with 'import substituting industrialisation (ISI)'.3 This confusion is apparent also in the paragraph in which Kaplan and Lewis (1996:123) refer to my comparison of the import substitution indexes of South Africa and Mexico in the period 1927-87.4 Questioning whether export expansion alone will be sufficient, and raising the possible need for reduced dependence on imports, obviously do not imply that export expansion should be neglected, and that there should be a 'more or less exclusive focus on the domestic market'. Nothing I say in my assessment of the ISP warrants the drawing of any such inference. My position is that we must achieve as rapid a rate of export growth as we can, and so far as possible rely on export expansion, but insofar as this, together with sustainable capital inflows, is not enough to maintain equilibrium in the balance of payments at some desired target growth rate of, say, 5 per cent a year, we would have to reduce our dependence on imports, through import substitution. In other words, import substitution would be used only to the extent that exports old or new, meet with inelastic world demands. This does not involve putting the entire burden of economic restructuring on import substitution, which it can be presumed would be less efficient. Kaplan and Lewis (1996:122) state that I am 'much more upbeat about the prospects for import substitution' than for export expansion. In fact I am 'upbeat' about neither. I state (1996:20) that, whereas South Africa came to 'the end of the easy stage of export expansion... comparatively recently', it came to 'the end of the easy stage of import substitution... some considerable time ago'. I therefore remark on our 'tough choice between difficult export expansion and difficult import substitution'. I state the obstacles to further import substitution in no uncertain terms. Indeed, though calling my view 'upbeat', they make much of my statement (1996:24) that 'significant new import replacement may initially 50 TRANSFORMATION 31 (1996) BELL DEBATE require considerable resources, and much resolve', which they (1995:123) describe as an 'ominous warning' which 'says it all'. They go on to quote my statement (1996:24) that if 'any of these conditions is lacking we may well, as the ISP report implies, have to rely more or less exclusively on export expansion'. They leave it at that, apparently under the impression that in the face of these obstacles their view about exclusive reliance on exports must prevail. I actually go on to say: 'In that case, however, we shall probably have to lower our sights considerably so far as improvements in ... growth performance... are concerned'. That is the really 'ominous warning', not the bit about 'resources' and 'resolve'. In any economy in serious difficulties, as ours is, when market forces alone cannot be depended upon to overcome them, practically any governmental strategy to deal with the problem is bound to require considerable resources and much resolve. The only economic question then is which policy or combination of policies can achieve any desired outcome most cheaply. In the present case, the overriding need is for increased production of tradable goods, whether exportables or importables. The extent to which this increase should consist of exportables and importables respectively depends simply on the respective costs at the margin of increasing production of each by, say, one US dollar. If the entire required increase in the production of tradables can be obtained more cheaply through export promotion alone, then the case for any promotion of import substitution falls away altogether. There can be no presumption, however, that the marginal costs of export promotion are smaller than those of import substitution, regardless of the lengths to which each is carried. As said above, we have a 'tough choice'. Export promotion is anything but costless.5 Despite the substantial costs which have been incurred, it seems that world demands for South Africa's exports are not proving sufficiently elastic, and that considerably more resources will have to be devoted to export promotion if we are to rely exclusively on exports. So far as World Trade Organisation (WTO) constraints on protection in the domestic market are concerned, all of the examples of export-promotion measures noted above are now no less subject to WTO constraints. In the case of both export promotion and import substitution, therefore, there is the problem of exploring the room to manoeuvre within WTO constraints. Given that neither the promotion of exports nor of import substitution may yield an immediate payoff, both would initially require not only considerable resources but much resolve. Importantly, however, in the case of the promotion of import substitution, much additional resolve would be required to overcome TRANSFORMATION 31 (1996) 51 DEBATE BELL the current irrational prejudice against it. The application of resources and resolve to the promotion not only of export promotion but also import substitution may nevertheless be excellent investments for the economy as a whole for the longer term. 'By way of conclusion: what space for industrial strategy?' In this section, Kaplan and Lewis state (1996:126-7) that for me 'macro-economic policy is all important', and they refer to my 'insistence on the absolute primacy of macro-economic policy'. Nowhere in my paper do I say any such thing, or indeed even use the term 'macro-economic policy'. Nor is any such view implicit in my argument. Conventionally, macro-economic policy primarily encompasses monetary and fiscal policy, neither of which is mentioned in the paper. The policy focus proposed in my assessment of the ISP report, and spelled out further, both above in the preceding section, and below falls within the domain of trade and industrial policy, albeit industrial policy conceived rather differently from Kaplan and Lewis. They have apparently, without justification, inferred from emphasis (on page 12, for instance), on the importance of 'macro-economic conditions', 'macro-economic factors', and a 'macro-economic perspective', that I see 'macro-economic policies' as the answer to our problems. The policy focus which I propose is aimed at overcoming a structural problem in the current account of the balance of payments and alleviating the foreign exchange constraint on South Africa's GDP growth rate. This is an aspect of maintaining macro-economic equilibrium, but it does not make the proposed policy focus macro-economic policy. What 'space' then is there for industrial strategy? As in the ISP report (1995:115), there is much talk now of the need for 'industrial restructuring'. But what does this mean? What has given rise to this need for 'restructuring'? What is the nature of the required restructuring? Does it simply mean making domestic producers more efficient and hence more competitive internationally? Or does it mean something more focused and concrete? These are the central issues for South African industrial policy, which need to be properly addressed. The view implicit in the ISP report is that the failure to increase productivity more rapidly underlies the problems of South African industry, and that overcoming them mainly involves raising productivity. There are however, alternative perspectives on both the causes of these problems, and on the nature of the process of restructuring required to overcome them. Underlying my argument is the view that, due to a structural problem in the current account, foreign exchange is an important effective constraint on the growth of the South African economy, and that alleviation of this constraint is a 52 TRANSFORMATION 31 (1996) BELL DEBATE necessary condition for a substantial, sustained improvement in the growth performance of manufacturing industry and the South African economy as a whole. This therefore should be a major focus for South African policy. This perspective implies that successful restructuring requires a change in the structure of the economy which makes some desired higher growth rate, of say 5 per cent a year, compatible with balance of payments equilibrium. This provides a very particular, concrete objective or focus for South African industrial policy, and a criterion for selecting industries to be targeted by the authorities. I do not argue that micro-level supply-side measures such as those proposed by the ISP, are not worth pursuing. They may be essential for the longer term. There is thus 'space' for ISP-type industrial policies. However, they do not deal with the fundamental, immediate problem. The ISP's more or less exclusive emphasis on raising productivity creates the danger that attention will be diverted from higher priority problems, and thus result in an inappropriate policy emphasis. It is this danger that my paper aimed at countering. The perspective which I have suggested illustrates the necessity of engaging properly in debate at this level if we are to define the required process of restructuring, and hence formulate industrial policy. There is no sign of any appreciation of this in the ISP report. From the outset the ISP report sees its business as producing policies aimed directly at raising productivity, by means of their supply-side measures. They, and now Kaplan and Lewis, present us with an almost purely technology-driven view of South African development. Singh, which has since come to hand, states well the essential point I was making in my discussion of the ISP report. Singh (1995:101) argues that in the Japanese case 'during the period 1950-73 a very important proximate goal of MITI was to ensure a current account balance at as high a growth rate as possible. In other words, the balance of payments was seen as the main constraint on fast economic growth in this period ... The government pursued this objective by a wide range of measures including, inter alia, a policy of extensive import controls, together with the promotion of certain key industries, which changed over time'. As Singh and Rodrik (1995) together make clear, this also applies to Korea and Taiwan. Singh (1995:101) rejects the World Bank definition of 'industrial policies ... as government efforts to alter industrial structure to promote productivity-based growth', as policy only to 'upgrade industrial structure', which (1995:102) confines its 'attention only to the supply side effects of productivity growth and technical change, as predicted by the TFP approach ...'.He says that 'to the extent that industrial policy helps to relieve the balance of payments constraint, most TRANSFORMATION 31 (1996) 53 DEBATE BELL sectors will benefit from higher rates of growth of production and hence productivity ... '. Hence, Singh contends that in assessing Japanese and Korean industrial policy 'one needs to look at the costs and benefits of industrial policy intervention in terms of their relieving the balance-of-payments constraint in the short and long run. More specifically, it would require, inter alia, an examination of the contribution of the favoured sectors to the growth of exports or to the reduction in the growth of imports over time'. This in essence is the view I was taking in my assessment of the ISP. It represents the 'credible policy implications that flow from the analysis' that Kaplan and Lewis (1996:1) say is absent, but is clearly present, in my paper. It does not represent, 'the counsel of despair'. Without such a focus, there is in my view no real hope for a substantial, sustained improvement in the growth performance of South African manufacturing industry and of the economy as a whole. As noted in my paper, the ISP was closely associated with Cosatu and the ANC during the course of their research, and several members of the research team are now in key policy-making positions. In this sense, it is true that, as Kaplan and Lewis (1996:1) tell us, the ISP inhabits 'the real world of policy research'. It is precisely for this reason that both the ISP report and the response of Kaplan and Lewis are so disturbing. NOTES 1 ' ™ O I l d i n g t 0 W r i g h t ( ' "3:58-60) when unadjusted for capacity utilisation, manufacturing rFP fell at 0.5 per cent a year in 1981-90, but at 0.06 per cent a year taking capacity utilisation into account. Moll (1990), as Kaplan and Lewis note, gets a negative TFP growth w ' ^ f U9 ** ceatayealfor 1981-90. While Moll uses the same method of analysis as wnght, his data set is different and, unlike Wright, Moll makes no adjustment for capacity uUisation. BelU et al (1992) gets a positive TFP growth rate of 0.55 per cent for 1983-90 • L ^ " * S a m e d a t a ** " w r i 8 h t - b u t a dlfl«rcnt method of analysis. Their choice of period, 1983-90 rather than 1981-90, however, is probably also crucial. There is thus no ctearcut evidence for the claim that there was 'strong' negative TFP growth in manufacturing industry in the 1980s. 2 S dingS ° f Y ° U n g ( 1 9 9 4 ) fly i n te f3CC o f widely-held views, including it seems those of . They will be challenged, perhaps successfully. They are the outcome, however, of a carernl cross-national growth-accounting exercise. They cannot be disposed of, as Kaplan and Lewis seem to think, in a couple of sentences involving a comparison of the United states with Germany and Japan; by inter-temporal evidence for Taiwan alone- or by extolling the achievements of Korea at the 'micro firm level'. v ' " / p a r t i c u l a r k i n d o f trade reg"1*5-in w h i c h the overall incentive structure is strongly biased towards production for the domestic market. It is generally contrasted with EOI which is also a trade regime, but one in which the incentive structure is more neutral as 54 TRANSFORMATION 31 (1996) BELL DEBATE between production for domestic and export markets. By contrast with ISI, import substitution is a process, involving local production of hitherto imported items, not a trade regime. Some measure of selective promotion of import substitution is quite compatible with EOI, as for instance, my reference (1996:18) to the experience of Korea and Brazil illustrates. 4. They say (1995:123): 'By disputing the conclusion that ISI has' been taken to excessive lengths, which make further import substitution exceptionally difficult', (Bell 1995:22) ... seems to suggest that further ISI is indeed possible'; and describe this as a 'backhanded endorsement of further ISI'. What I actually say is: 'This also does not suggest that the process [of import substitution] has, for peculiarly South African political reasons, been taken to excessive lengths, which make further import substitution exceptionally difficult'. The omission of the first half of the sentence changes its meaning. Apart from their omission of the crucial reference to 'political reasons', they clearly regard the 'process [of import substitution]' as interchangeable with ISI. Nowhere on page 22 do I refer to 'ISI'. Furthermore, to imply, as Kaplan and Lewis do, that further import substitution is not possible is absurd - the only relevant question is at what cost. Finally, rising import-penetration ratios, noted by Kaplan and Lewis in the last two sentences of their paragraph are not necessarily indicative of limited capacity for further import substitution. 5. To give just three examples: The General Export Incentive Scheme (GEIS) has required considerable resources, equal to an average of over 6 per cent of the value of all exports eligible for subsidisation under the scheme. The various large-scale '37(E)' projects, aimed largely at increasing exports, also involved considerable resources in terms of forgone tax revenue. And, under the Motor Industry Development Programme, exports of automotive components at the moment effectively receive a subsidy, in the form of rebated import duty, equal to 46 per cent of their local content value, and there are in addition fears that this is having damaging effects on the domestic assembly and component industries. REFERENCES Bell, T (1995) 'Improving manufacturing performance in South Africa: a contrary view', Transformation, 28. Bell, T (1993) 'Should South Africa further liberalise it's foreign trade?', in M Lipton and C Simkins (eds) State and Market in Post-Apartheid South Africa, Johannesburg, Witwatersrand University Press. Bell, T and N Cattaneo (1996) 'Foreign trade and employment in South African manufacturing industry', International Labour Office, Geneva (forthcoming). Belli, P, M Finger and A Ballivan (1993) South Africa: A Review of Trade Policies, The World Bank Southern Africa Department, Informal Discussion Papers on Aspects of the Economy of South Africa, Discussion Paper 4. Gereffi, G (1990) 'Paths of industrialisation: an overview', in G Gereffi and D Wyman (eds), Manufacturing Miracles: Paths of Industrialisation in Latin America and East Asia, Princeton, New Jersey: Princeton University Press. Helleiner, GK (1994) Trade Policy and Industrialisation in Turbulent Times, London and New York: Routledge. Joffe, A, D Kaplan, R Kaplinsky and D Lewis (1995) Improving Manufacturing Performance in South Africa: Report of the Industrial Strategy Project, University of Cape Town Press. TRANSFORMATION 31 (1996) 55 DEBATE BELL Kaplan, D and D Lewis (1996) 'Improving productivity performance in South Africa: a response to Trevor Bell', Transformation, 30. Kaplinsky, R (1995) "Capital intensity in South African manufacturing and unemployment, 1972-1990, World Development, 23(2). Kim, J-I and LJ Lau (1992) 'The source of economic growth of the newly industrialised countries on the Pacific Rim', unpublished manuscript, Stanford University. Moll, T (1990) 'Output and productivity trends in South Africa: apartheid and economic growth', unpublished PhD dissertation, University of Cambridge. National Productivity Institute (1990) Productivity Focus, Pretoria, NPI. Ocampo, JA (1994) 'Trade Policy and Industrialisation in Colombia, 19676-91', in Helleiner (ed). Rodrik, D (1995) 'Getting interventions right: how Taiwan and Korea grew rich'. Economic Policy, 20. Singh, A (1995) 'The causes of fast economic growth in East Asia', UNCTAD Review 1995 UNCTAD, Geneva. Wright, A (1993) 'Productivity growth in the South African manufacturing sector: deriving and assessing neoclassical TFP measures' in K Kruger (ed), Papers submitted for the Competition to Promote Academic Research into Industrial Policy and Development in South Africa, Industrial Development Corporation, Sandton. Young, A (1994) 'Lessons from the East-Asian NICs: a contrarian view', European Economic Review, 38. Young, A (1994a) 'The tyranny of numbers: confronting the statistical realities of the East Asian growth experience', unpublished paper. 56 TRANSFORMATION 31 (1996)