Article Class, Distribution and Redistribution in Post-Apartheid South Africa Jeremy Seekings andNicoli Nattrass Soon after democratisation in 1994, the last vestiges of statutory racial discrimination were removed in South Africa. But the new government inherited a society where inequality could not be reduced to race alone and inequality persisted in the face of formal political equality. The best available data suggests that the Gini coefficient remained stable or even edged up during the 1990s, as inter-racial inequality declined but intra-racial inequality rose sharply. Why has the demise of apartheid and the onset of democracy not been accompanied by a decline in inequality? As Aristotle noted two and a half millenia ago, democracy entails rule by the poor because the poor constitute the majority. Why hasn't the formal equality of representative democracy induced the ANC to introduce more effective pro-poor reforms? Regrettably there are no careful studies of policy-making in the post-apartheid state corresponding to (say) Posel's study of influx control under the apartheid state (Posel 1991). In the absence of such studies, we have to resort to the broad sweep of political economy, the strength of which is the linkage of politics to economics and the weakness of which is its treatment of the state as a 'blackbox'. The literature on the political economy of post-apartheid South African can be divided into three main approaches. The first approach, represented most influentially in the work of Bond (2000) and Marais (1998, 2001), analyses South Africa as a battleground between capital (especially international capital and its supposed agents, the IMF and World Bank) and the largely undifferentiated South African masses. Whilst neo-Marxist in its critique of capitalism, this approach lacks any empirical analysis ofthe class structure. Its empirical content comprises primarily a critical analysis of the TRANSFORMATION 50 (2002) ISSN 0258-7696 Jaiwny SaeMng* and Nicotl Nattras* ANC's macro-economic, housing and urban infrastructural policies, and most recently the government's commitment to privatisation. Its political analysis does not extend far beyond the assertion that the ANC leadership has 'sold out'. A different approach is evident in the work of scholars with close links to the labourmovement, including Adler and Webster (1999,2000; also Gelb and Webster 1996, Webster 2001) and Baskin (2000). These scholars have to explain the continued participation of COSATU in an alliance with the ANC. For them, post-apartheid policy reflects a class compromise, in which pro-capitalist macro-economic and other policies are weighed up against labour's achievements in terms of labour legislation (especially the Labour Relations Act) and corporatist institutions (especially NEDLAC). Scholars in this tradition tend to see society very much in terms of two classes that correspond to the 'two nations' identified by President Mbeki: one rich and white, the other poor and black, both largely undifferentiated (Adler and O'Sullivanl996). Our own work falls into a third approach in the literature on political economy. We recognise that capital retains enormous power, not because of the voting power ofcapitalists in elections but because oftheir importance to the economy. In an era of globalisation, in which capital is more mobile and can exercise more readily the 'exit option' outof a country (or simply not enter in the first place), the power of capital relative to both other classes and national governments is surely greater than before (see Nattrass 1999). But, we have argued, it is simplistic to view a society like South Africa simply in terms of capital and labour. There remain important divisions and conflicts of interest within what other scholars refer to as the 'working class as a whole' (Nattrass and Seekings 1998c, Seekings2000). Moreover, social and economic change has eroded greatly the correlation between race and class in South Africa. High levels of inequality are increasingly based on intra- racial not inter-racial inequalities (Nattrass and Seekings 2001b). The class structure ofpost-apartheid South Africa comprises three major groups of classes in South African society (see figure 1). At the top is an increasingly multi-racial upper class or elite. In the middle lie workers in a range of classes: the 'semi-professional' class (teachers and nurses), the 'intermediate' class (ie white-collar workers in public and private sectors) and most of the 'core' or urban industrial working class. At the bottom are the marginalised sections of the working-class (including especially farm and domestic workers and their dependants) and households where no one Class, Distribution and Redistribution in Post-Apartheid SA is in employment (see further Seekings and Nattrass, forthcoming: chapters 7 and 8). Figure 1: The Post-Apartheid Class Structure Mean %of %of household all all income house- income (RAnonth holds / A \ * Upperclass 1993) 7020 12 45 / \ Sara-prof. 3264 5 / \ class 2257 19 / \ ^ Intermediate 45 / \ class 1187 19 / \ Corewoddng / \ class 1442 5 / \ Pedytraders / V 1 Marginal 618 12 / \ working class 10 Underclass and 413 28 other Notes: Data is from the 1993 SALDRU survey. The 'upper class' comprises households headed by people in managerial, technical or professional occupations, or with substantial income from assets or entrepreneurial activities; the 'semi-professional class' comprises households headed by teachers or nurses; the 'intermediate class' comprises households headed by routine white-collar, skilled or supervisory workers; the 'core working-class' comprises households headed by semi-skilled or unskilled workers outside of agriculture and domestic work; the 'marginal working class' comprises farm and domestic workers; the 'underclass and other' category comprises households with no members in employment and negligible income from entrepreneurial activity or assets. The post-apartheid 'distributional regime' has its roots in the distributional regime under apartheid, which itselfreflected the class interests of powerful white constituencies (Nattrass and Seekings 1997). The basic framework of apartheid labour market and welfare policies was designed to protect the interests of sections of the white working class in the 1930s. Massive investment in public education for white children in the 1950s and 1960s Jeremy Seeking* and Nlcoli Nattrass resulted in white workers securing the skills that enabled them, in the 1970s and 1980s, to commandhigh incomes in free labour markets, largely removing their dependence on direct state interventions (such as job reservation through the 'colour bar') and even on the institutional framework for wage determination. By this time, however, the state' s labour market and welfare policies had begun to serve the interests ofthe newly skilled African working class, which effectively opposed the dismantling of a range of pro-labour policies. Policies that had protected the incomes of white workers from the 1920s until the 1970s now served to protect the incomes of semi-skilled and some unskilled African workers. To the extent that this encouraged job- shedding and capital-intensive growth, these policies were at the expense of other unskilled workers and the growing ranks of the unemployed. Today, the organised African working class enjoys household incomes above the median but below the mean in South Africa. Whilst the class is clearly 'exploited' in some senses, and receives incomes that are low by comparison with South Africa's privileged elite, it also seeks to protect its semi-privileged position. It resists reforms to labour market and other policies that would steer the economy down a more labour-absorbing growth path (see Nattrass 1999,2000a, 2000b, 2001), and resists any extension of the tax base that might transform them from net beneficiaries of fiscal redistribution (as they are at present) into net contributors. Compared to capital and organised labour, the poor have little leverage over policy-making. They cannot threaten to withhold their capital or their labour. But they are numerous, giving them potential electoral strength. Elsewhere (Nattrass and Seekings 2001 a) we identify three reasons why the poor use so feebly this potential electoral strength. First, the evidence suggests - contra the conventional wisdom - that the poor have limited expectations of material gains in the short-term, because they are both patient and they recognise the constraints on the state (see further Charney 1995, Nattrass and Seekings 1998b). Secondly, poor voters in most parts of the country do not have the option of switching their support to pro-poor parties because either there are no such parties in the area or, if there are, they lack credibility. Thirdly, and perhaps most important of all, some policies are more opaque than others. Poor voters may be very sensitive to the delivery of services (such as public education, health care and welfare) that affect them directly. But it is not easy for voters to discern precisely the extent to which the government is responsible for the shortage of employment opportunities for the unemployed. Thus, whilst the poor express criticism of Class, Distribution and Redistribution in Post-Apartheid SA union action that is seen to impede job creation and may engage in direct action against immigrants who are seen to be taking scarce jobs, they may not punish the government for the lack of job creation. In the long-term, electoral self-interest might push the ANC toward more egalitarian policies, but pressures are weak in the short-term. The result of these political forces is a set of government policies that offers something to everyone in a giant but precarious political compromise. Those groups with an interest in free markets get conservative macro- economic policies. The industrial working class gets protective labour market legislation, a favourable deal from the budget (especially through the continued restriction of the tax base), and have (to date) forestalled major privatisation. Many public sector workers (including, especially, teachers) continue to enjoy favourable salaries. The poor get a certain amount of redistribution through the budget (including especially an old-age pension system whose coverage and generosity is unique in the developing world). Racial rhetoric (including the 'two nations' thesis) helps the ANC to preserve its multi-class support base. If government performance is poor, the ANC argues, this is due to the apartheid legacy and beyond the ANC's control. We argue that the legacy of the past is heavy, but that the government has chosen not to tackle it fully - because of the political pressures on the government itself. In this paper we examine, first, the distributive consequences of post- apartheid South Africa's economic growth path and of more obviously redistributive policies. We then turn to evidence on processes of class formation and changing patterns of income distribution. Finally, we identify a series of policy reforms or strategic emphases that would have more egalitarian effects in South Africa's labour-surplus economy. The Growth Path and Redistribution Unemployment is a major determinant of poverty and inequality, and there was clear evidence of high and rising unemployment in the 1990s (Nattrass 2000b). At the same time, the structure of South African employment has become more skilled (Bhorat and Hodge 1999). This is the result of two trends: a general shift away from unskilled labour in all sectors and the especially sharp decline in the labour-intensive mining sector. Employment shedding off commercial farms contributed to the shift away from unskilled employment in South Africa over the decade (Simbi and Aliber 2000:7). Jeremy Soefcings and NkxHi Nattrass Figure 2 illustrates the dramatic decline in employment in mining and agriculture across the 1990s - especially with regard to less-skilled workers in those sectors. Thus, even in the historically unskilled labour intensive mining sector, there has been a shift towards higher paid, better skilled workers - but at the cost of employment overall. Figure 2. The Decline in Employment in Mining and Agriculture (Sources: South African Reserve Bank Data and data extrapolated from Edwards (2000)) -Mining employment (1990=100) —Agricultural employment 1 (1990=100) • Least skilled mining and agricultural employment Real earnings per worker increased over the period (see Figure 3), but especially in the late 1990s - being driven mainly by increases in the public sector. Rising average earnings are a function of two factors: trade union pressure (especially in the public sector) and the changing skill composition of the employed workforce. Given that unskilled workers are bearing the brunt of retrenchments, one would have expected average earnings to have risen (as better-paid, relatively skilled workers remained in employment). One would thus expect labour productivity to be rising as well. As shown in Figure 3, this indeed appears to be the case: labour productivity has risen as employment declined. This outcome is consistent with the goals and policies of the government's 'high productivity now' (HPN) growth strategy (Nattrass 2001). Notice that the growth in labour productivity exceeded that of average wages for most of the period - thus facilitating a slight growth in the profit share. This suggests that the growthpath has been relatively kind to employed (especially skilled) labour and capital, and unkind to unskilled labour and the unemployed. Class, Distribution and Redistribution in Post-Apartheid SA Figure 3. Labour productivity, employment, average wages and the profit share (indexes, 1990=100) - Total non- 140 agricultural employment •Labour productivity •Total real remuneration per worker . index of gross profit share This is not to suggest that the growth path has heen good for capital accumulation. To the contrary, economic growth has been lacklustre andper capita incomes have fallen across the 1990s. But some economic activities have suffered less than others - and have even benefited from the policy interventions and institutional forces acting on the economy to drive it 'up the value chain'. In other words, the HPN growth strategy has paid dividends for some workers and some capitalists, but it has done nothing to improve the labour absorption capacity of the economy. The gap between labour market insiders and outsiders (especially the unemployed) is thus likely to remain large (ifnot widen) and there appear to be few forces acting to narrow overall inequality in South Africa. Has redistribution through the budget ameliorated the inegalitarian impact of the growth path? The scope for increased redistribution through the budget in South Africa after 1994 was constrained by three factors. The first was the government's commitment to conservative macro-economic policies, which precluded large increases in overall government spending without matching increases in taxation. Taxes rose for just about everybody in the mid-1990s (Simkins and Woolard 2000), but this was relayed into reduced budget deficits rather than increased spending. The second was the fact that the government inherited, in 1994, a budget that was already significantly redistributive. Thirdly, the government remained subject to political pressures that diverted it from the policy reforms (particularly in the labour market) that would have reduced inequality most markedly (by prioritising job creation). Jeremy Seekings and Ntedi Natttass The development of the government's macro-economic strategy has been well documented (see eg Marais 1998,2001, Michie and Padayachee 1998, Habib andPadavachee 2000), but recentfiscalincidence analyses have not been acknowledge in the political-economic literature. McGrath et al (1997) have shown how redistributive the budget was at the end of the apartheid period, reducing the Gini coefficient from about 0.7 (for gross or original income) to about 0.6 (for post-tax/transfer income, taking into account also the value of' in kind' benefits from public education and health care). Redistribution happened because the top quintile received much fewer transfers or in kind benefits than they paid in tax, whilst everyone else received more benefits than they paid taxes. Van der Berg (2000, 2001) argues that the budget became still more redistributive after 1994. Taking into account government spending on welfare transfers, public education, public health, subsidies for housing and capital expenditure on the provision of water, he estimates that spending on the poorest 40 per cent of households (ie quintiles 1 and 2) rose by about 5 0 per cent in real terms between 1993 and 1997. A small part ofthis was made possible by reduced spending on the rich, in that spending per capita on the top quintile actually declined. But the lion's share of extra spending on the poor in the mid- 1990s came from increased and well targeted spending by the government. This targeted expenditure entailed not cash income, in the form of government welfare transfers, but rather benefits in kind — especially in terms of public education. Van der Berg shows that the removal of indirect discrimination in teachers' salaries, together with the provision of some extra teachers and hence reduction in pupil-teacher ratios, entailed massive increases in spending in 'African' schools, ie schools with overwhelmingly African students, especially in poor, rural areas. Under apartheid, teachers were paid on different salary scales. In 1996-7 all teachers were moved onto a single, consolidated salary scale, based on the scale of the former white education departments. Approximately 40 per cent ofteachers were moved into higher salary brackets, and average salaries rose by between 12 and 15 per cent (RSA 2000a:4.5). This shift was probably driven by the need to deracialise salaries, especially given pressure from African teachers, rather than a concern with the poor. But if the value ofpublic education is deemed equal to the cost of providing it, then the poor can be said to have benefited substantiallyfromthis increase. Van der Berg (2000) calculates that the share of total spending on public education accruing to the lowest income quintile rose from 23 per cent to 29 per cent between 1993 and 1997. Class, Distribution and Redistribution in Post-Apartheid SA Describing an increase in teachers' salaries as 'increased spending on the poor' is somewhat misleading-unless, of course, the increased salaries are accompanied by improved quality of teaching. However, teachers in schools in poor areas remain inadequately qualified, often poorly motivated, and in too many cases simply incompetent. It is probably fair to conclude that, at least in the short-term, the major beneficiaries of increased educational spending were teachers (who are not relatively poor), not the students sitting in their classes. In other areas of public spending the gains to the poor were also ambiguous. The health, housing and infrastructural budgets may have become better targeted on the poor, but it is unclear how much the poor actually benefited in terms of the quality of the services provided. If the evidence on benefits 'in kind' is ambiguous, the evidence on cash benefits actually undermines the government's claims. The single most important instrument of direct redistribution through the budget is the old age pension. Between 1993 and 2000 the real value of the old age pension declined by a total of about 20 per cent. The fact that spending on old age pensions declined faster than total government spending indicates'that the elderly and their dependants bore the brunt of the adjustment in this regard. Similarly, the savagery ofthe 1998 reform ofpublic financial support for low- income single parents suggests that the government is passing a significant portion of the burden of budgetary austerity onto the poor. Cuts in welfare benefits have been defended on the basis that social welfare is to be 'developmental' rather than dispensing 'handouts' (see discussion in Budlender 2000:125,130-3). But what this means in practice is unclear, and in the current context of rising unemployment and sluggish growth, such an approach imposes costs on the poor (at least in the short- term). In essence, it is probably fair to conclude that six year's-into the post- apartheid era, the distributional regime remains much the same as ithad been during the late apartheid era. Growthpath policies and labour market policies buttress the earnings of the middle class (including in the public sector) and the industrial working class, whilst eroding the demand for unskilled labour and reducing the prospects that the unemployed would secure employment. The welfare system provides generous assistance to old-age pensioners, limited assistance to single parents, and very heavily subsidised public services for children especially - but no assistance to the bulk of the poor, whose poverty was due to the lack of job opportunities. Jeremy Seekings and NicoJi Nattrass The divide that existed between white and African households under the early apartheid period continues to shift, separating growing numbers of better off African households from the African poor. The proportion of African households resorting to the private sector for the provision of welfare has grown steadily. By 1997 almost one-fifth of the population was covered by private medical aid schemes, and coverage was closely related to income (South African Health Review 1997:82). Most formal sector employees—including the core working and intermediate classes — are now covered by private sector retirement funds. Growing numbers of African people in urban areas now send their children to schools that are private or semi-private (in that, whilst in the public sector, they charge high fees to provide a superior education). A multi-tier system of education, health and retirement pensions has emerged, with access to the higher tiers dependent on prior employment. As was the case with the apartheid distributional regime, the public welfare system makes no provision for the many poor people who are not old enough for the pension nor young enough to qualify for child support. Despite the limited reform of unemployment insurance (which extended coverage to workers in previously excluded sectors, such as domestic labour, and provided higher proportional benefits for low-income workers than high-income workers) there is still no provision for the long-term unemployed, nor for people who have never been employed. Whereas the apartheid distributional regime was premised on full employment, the post-apartheid distributional regime operates in the context of extremely high unemployment (see Nattrass and Seekings 1997). The absence of any welfare net for the unemployed thus constitutes a major problem. The semi-privileged position ofpolitically powerful African groups - including the urban, industrial working class, sections of the intermediate class and the semi-professional class (especially teachers) - gives them good reasons to oppose a universal welfare system in that radical welfare reform would require increased taxation on them, making them subsidisers of the poor rather than the beneficiaries of redistribution from the rich. Perversely, therefore, the legacy of apartheid included the formation of a politically powerful, cross-racial coalition of classes, some a lot more privileged than others, with an interest in opposing radical reforms to the distributional regime. 10 Class, Distribution and Redistribution in Post-Apartheid SA Class Formation and Inequality There is some evidence to suggest that inequality has grown in the 1990s. Whiteford and van Seventer (2000), using census data, show how the distribution of income became more unequal between 1991 and 1996. Inter- racial inequalities declined (although they remained large, of course), whilst intra-racial inequalities continued to grow. The incomes of the richest ten per cent of African households rose by 17 per cent, whilst the incomes of the poorest 40 per cent of African households fell by 21 per cent. Rich white households maintained stable incomes, but relatively poor white households saw a big absolute decline in their incomes (Whiteford and van Seventer 2000:14-19). The real winners in terms of rising incomes were the better off African households. Of the total real increase in income between 1991 and 1996,40 per cent went to the richest ten per cent of African people, and a total of 62,5 per cent went to the richest 40 per cent of African people. The poor majority of African people barely benefited at all (Whiteford and van Seventer 2000:1). This data can also be presented in terms of income categories. The share of total, ie national, income earned by African households with an income of at least R72 000 (in 1996 prices) rose fromnine per cent in 1991 to 14,5 per cent in 1996 (Whiteford and van Seventer 2000:22-4). These findings are broadly corroborated by data from the KwaZulu-Natal Income Dynamics Study (KIDS). KIDS found that the proportion of its panel with incomes below a fixed poverty line rose from 35 per cent in 1993 to42 per cent in 1998. At the same time, the proportion with high incomes also grew, indicating deepening polarisation (Carter and May 1999). These trends are consistent with our argument about the post-apartheid growth path: employed workers, particularly those with skills, are relatively advantaged by HPN policies; whereas less-skilled workers and the unemployed are disadvantaged (at least in the short- to medium-term). The gap between skilled and unskilled workers is thus likely to grow. Cichello et al' s (2000) analysis of changes in individual earnings in the KIDS data tends to support this. They show that the earnings ofworkers in regular employment rose by 37 per cent between 1993 and 1998, compared to an overall average change in earnings of just sevenper cent. The earnings ofworkers in regular employment grew faster than the average for everyone in the sample. Some ofthis spectacular increase was because new entrants into formal employment had higher wages than those who left. But even among workers who were in formal employment in both 1993 and 1998, earnings rose by 20 per cent. 11 Jeremy Seekings and Nicoli Mattrass Cichcllo ct al do not disaggregate the category 'regular employment' into discrete classes, but it is likely that the core working class and intermediate classes enjoyed positive real income growth during the late 1990s. This does not mean - as Cichello et al emphasise - that all members of these classes prospered. The earnings of workers who lost or left their jobs plummeted. The aggregate gains ofclasses such as the core working class and intermediate class should not obscure the fact that the composition of these classes shifted, as individuals and households dropped into lower classes, to be replaced by individuals and households that were upwardly mobile through the class structure. The Changing 'Middle Class' Has this slight widening of income inequality been accompanied by any major change in the class structure? Perhaps the most striking change in South African society in the 1990s has been the accelerated growth of what is generally called the black or African middle class. Whiteford and van Seventer's data shows that the lion's share of increased national income between 1991 and 1996 accrued to a small but rich minority ofAfrican people. The combination of improved access to education, the removal of any remaining restrictions on upward occupational mobility and the introduction of affirmative action policies by both the state and private sector meant rapidly expanding opportunities for some. This trend has presumably been exacerbated by the flight abroad ofmany skilled white workers, intensifying skills shortages. Under apartheid, as Crankshaw (1997) has shown, upward mobility among African people was largely limited to the semi-professional occupations of teaching and nursing and white-collar occupations entailing little authority. Managerial posts were largely limited to the bantustan bureaucracies, and there were very few African businessmen. The end of apartheid saw a major shift in the range of opportunities open to African people. The number of African employees in senior managerial positions in the public and parastatal sectors grew rapidly. African managers were also appointed to senior managerial positions in the private sector, especially in jobs where employers considered it important to have someone with a black face or good political connections. As universities turned out more and more black graduates, so the proportion of professional and managerial posts filled by African men and women rose. Most remarkably, there was a clear increase in the number and importance 12 Class, Distribution and Redistribution in Post-Apartheid SA of African businessmen. According to calculations by BusinessMap (1999:11,26-7), just one per cent of the market capitalisation on the Johannesburg Stock Exchange (JSE) was under black control in late 1995, but by January 1999 the share had risen to 5.5 per cent. This share was controlled by 3 5 listed companies. A further 11 per cent of JSE market capitalisation was described as being 'under black influence' (ie with black firms having a significant but not controlling interest); by far the largest of these was South African Breweries. Even in the aftermath of the 1998 stock market crash, this amounted to almost R60 billion. Lacking capital, emerging African capitalists were dependent on the financial institutions to buy into corporate ownership. The growth of the black corporate class in South Africa was thus based on the country's well-organised banking sector, active equity market and the fact that control and ownership were already largely local rather than multinational — factors that distinguished South Africa from most other African countries (Randall 1996). But it has meant that 'black economic empowerment' has been driven by debt (with consequences for risk and hence entrepreneurial decision-making). Black empowerment deals were concentrated in financial services (typically with black-owned firms forming partnerships with established institutions), information technology and telecommunications, and the media (including publishing). Penetration was more limited in the traditional industrial sectors, excepting the case of Johnnie. Whilst these emergent African capitalists were not directly dependent on the state, they were especially active in sectors or industries in which the state wields some control, for example through licensing (eg fishing and media) or influence, through tendering, procurement or privatisation policies. Some of the fastest-growing 'black' investors were the ten or so trade union investment companies. Because they are not listed on the JSE it is apparently not possible to provide a clear aggregate figure for their investments (BusinessMap 1999:65). There are not, of course, a huge number of African businessmen (or women). In no capitalist society do capitalists constitute more than a tiny proportion of the population. In terms of sheer numbers of people, it is the teachers, nurses, professionals and junior and middle managers who constitute the 'middle class'. In South Africa, there are still more white people than African people in the middle classes (although it is likely that white people comprised less than half of the top income quintile by the end of the 1990s). The rapid growth of the black middle class has been noted by the 13 Jeremy Seekings and Nicoli Nattrass marketing and market research industries. In 1998 the Bureau of Market Research at UNIS A conducted a survey on the lifestyle, beliefs and attitudes of the emerging African 'middle class' in Gauteng (defining households as 'middle class', somewhat idiosyncratically, if they 'had moved to a better dwelling or if there was an improvement in the occupation/employment of an adult household member over the past five years' (BMR 1999:1). Many had moved into formerly white middle class residential areas. Most were young, and had not iived in their present house for long. This class had quite distinctive consumer preferences. The product which was seen as conferring the highest status on people was the car—with men preferring BMWs or Mercedes, whilst women favoured VW Golfs or Jettas. Cellphones and clothing also conferred status. They preferred to buy clothes from Foschini, rating especially highly Polo shirts and Pierre Cardin suits. They aspired to shop at Woolworths. Over the past few years they had come to use financial institutions far more. The Standard Bank was the favourite bank, and Old Mutual and Mutual and Federal the favourite providers of long- and short-term insurance respectively. They took their holidays at inland resorts or, increasingly, overseas. A high proportion - almost 30 per cent - aspired to have their own businesses. Whilst the BMR investigation tends to treat the African middle class as a somewhat exotic species, it helps to fill the gap caused by the absence of detailed sociological or ethnographic studies of this new social group. Survey and other data is beginning to suggest that this class will reproduce itself overtime,ie that the privileges ofthe current generation will be passed onto their children. There is a clear relationship between educational attainment of children and the class ofthe household as a whole, even when inter-racial differences were removed from the picture. Figure 4 shows the mean grade attainment of African children according to some of the classes defined as in Figure 1. The African middle class is clearly growing rapidly, but is there any evidence that the class structure as a whole is changing? Is it perhaps rather that the racial composition of different classes is changing? Whiteford and van Seventer's calculations show that the number of households in the 'middle class', defined simply in terms of income as households with an income(in 1996prices) ofover R72000peryear,grewbetwecn 1991 and 1996, but not as fast as the population as a whole. The middle class therefore actually shrank in relative terms. The African middle class grew rapidly, whilst the white middle class shrank very slightly. Such trends are consistent 14 Class, Distribution and Redistribution in Post-Apartheid SA with the hypothesis that the overall class structure changed little, although the racial composition of the upper income group was shifting rapidly. highest school grade completed, bycurrentag* and dais, African children only, 1983 SPC 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Age Note: data from 1993 S ALDRU survey. UC is upper class; SPC is semi- professional class; IC is intermediate class; CWC is core working class and MWC is marginal working class; data for other classes is not reported here. Further research is required into the changing nature of the class structure, using data from surveys conducted after 1993. But the available data on incomes suggests that the biggest 'winners', as Whiteford and van Seventer put it, have been better-off African people who have moved into and through 15 Jeremy Ssekings and Nicdi Nattress the 'middle class'. The losers, of course, have been those people without regular employment at all, many ofwhom do nothave the skills to secure jobs in a skill-intensive economy even if the growth rate picked up remarkably. Class inequalities show little sign of abating, even if the upward mobility of aminority ofAfrican people means that inter-racial inequalities have dropped markedly. The 'Re-segmentation' of the 'Working Class' A number of scholars, together with the trade union movement, point to new processes of segmentation within the labour market, between those in regular forms of employment and those in non-regular forms such as part- time, temporary and sub-contracted (out-sourced) employment. Elsewhere we argue that employers' recruitment strategies favoured some job-seekers over others, and suggested a large percentage of the unemployed should be considered as members of an 'underclass' on the grounds of the disadvantages they suffer in terms of access to employment options (Seekings and Nattrass forthcoming:chapter 8). Should we reassess our class categories, and in particular the boundaries between core working class, marginal working class and underclass, in light of the shift from regular to non-regular forms of employment? Kenny and Webster, in an analysis of the growth of casualisation and sub-contracting, conclude that: ... profound changes are taking place in the labour market: new winners and new losers are emerging. Above all, the resegmentation of the labour market is entrenching the position ofthe 'old losers' — the unemployed, informal sector, and migrant workers. Flexibilization is creating an insecure, lower paid, and unprotected workforce. (1998:240) Kenny and Webster attribute the shift to non-regular forms of employment to employers' search for lower labour costs in the context of globalisation and 'neo-liberal' government policy. They pay little attention to the role of labour legislation and other government policies that cannot be described as typically 'neo-liberal'. As we have argued elsewhere, various government labour market policies have the direct or indirect effect of pushing up the relative cost of unskilled labour (see eg Nattrass 2000a). It is not surprising that the shift to non-regular forms of employment has been especially marked in labour-intensive sectors, including cleaning, security, and the retail sector. Kenny and Webster acknowledge, but barely explore, the role of unions in accentuating divisions in the workplace, as they protect the interests of permanent workers against those of casual workers. 18 Class, Distribution and Redistribution in Post-Apartheid SA Theron (2000) argues that we need to distinguish between casualisation and out-sourcing as the latter need not entail the former. However, in practice, sub-contracting often entails a decline in hourly wages, benefits and employment security. Some of the workers who are shifted from regular to non-regular forms of employment have skills and may realise steady or even increased earnings. But it is likely that the bulk of newly casualised jobs involve the less skilled in occupations that we classified as core working class. Does casualisation or out-sourcing change their class position? There is an argument that such workers should be classified in the marginal working class. In other words, changes in the labour market are such that marginality should not be defined in terms of sector only (with agriculture and domestic work qualifying as marginal), but also in terms of the form of employment. If we were to do this, then it is likely that the marginal working class is growing and the core working class is shrinking. Whether or not we interpret this shift as a changing class position, the extent of casualisation, out-sourcing and sub-contracting is unlikely to be large relative to the very high and apparently, rising rate of unemployment in South Africa. The marginalised sections of society are dominated by the long-term unemployed, especially those in the 'underclass' with poor access to vacancies and hence little chance of securing employment. Towards a Social Democratic Agenda Is there anything that government can do to reduce inequality whilst ensuring a sustainable growth in income? In this concluding section we sketch the key components of what we consider to be a social democratic policy agenda for a middle-income labour-surplus economy like South Africa's. We describe the agenda as social democratic because it combines the goals, of improved welfare (in the broad sense of the standard of living of citizens), reduced inequality and extended democracy, to be achieved through a combination of state and market. (a) Full employment The most important element of a social democratic agenda in a massively labour-surplus economy must be the reduction of unemployment through sustained job creation. Sustained job creation requires mat the South African economy achieves a higher rate of growth and follows a labour- absorbing growth path. Public policies must work towards both of these goals. We suspect that the state could pursue a less contractionary - or more 17 Jeremy Seeking and Nicdi Nattrass expansionary - fiscal policy than it has under GEAR, and as critics in the labourmovement and elsewhere have demanded. At the same time, however, it is imperative that any labour market policies that impede low-wage job creation be reformed (see eg Nattrass 2000a, 2000b). The poor cannot wait for the benefits of a 'high-productivity now' growthpathto trickle down to them. This probably means tolerating greater wage inequality (as low-wage, labour-intensive employment expands), in order to reduce overall income inequalities by creating jobs for the unemployed. We propose the reform of legislation providing for the extension of centrally-bargained agreements,. increased scope for firm-level rather than industry-level bargaining, and the reform of the Employment Conditions Commission to ensure that account is taken ofthe employment effects ofwage-setting. On this, our views contrast starkly with the official views of the labour movement. We are not proposing the abandonment of employment protection or health and safety provisions for some workers and nor are we proposing wage cuts. Rather, entrepreneurs shouldbe allowed to create new employment at low wages if necessary. Average wages will probably fall, but as a result of job creation (which reduces inequality) not wage cuts (that probably increase it). This will help address the unemployment problem at the margin by nudging the economy in a more labour-intensive direction. It is only through rapid growth that unempleyment will be significantly reduced. Our position is a democratic one. Democracy and organisation should be extended into low-wage workplaces, not agreements bargained by other, self-interested parties. The extension of democracy into the workplace should surely entail allowing the people most affected by the consequences of a decision to play the leading role in making that decision. Wage agreements should, as far as possible, be made by those workers who stand to reap the benefits (higher wages) or suffer the consequences (unemployment) ofthe deal. In sectors where there is no collective bargaining, the Employment Conditions Commissions should be tasked with balancing the need for acceptable minimum wages with job creation for the unskilled and unemployed. The idea of employment subsidies has been raised, most notably by Sam Bowles (see eg Heintz and Bowles 1996) and introduced in a limited way in the 2001 budget. The idea is that employers will be paid a wage subsidy which willlower the cost of employment without reducing workers'take-home pay or aggregate demand in the economy. But the potential benefits need to be weighed up against the possible costs, which include subsidising inefficient 18 Class, Distribution and Redistribution In Post-Apartheid SA use of labour, administration costs, deadweight losses (ie subsidising employment which would have been created without the subsidy) and so on (see Standing etal 1996:458-60). Taxing capital assets to pay for the subsidy (as suggested by Heintz and Bowles) should encourage labour-intensive production at the level ofthe firm, but the economy-wide effects are unclear. The overall impact of financing the wage subsidy out of general taxation (as operationalised by the government) is even more unclear. More research is required on this issue. In an earlier debate, Gelb and Webster (1996) seemed to argue that the incomes ofthe 'working class as a whole' would be maximised through wage increases, even if there is a trade-off between employment and wages. This argument makes no allowance for distribution within the 'working-class' - by which Gelb and Webster clearly meant all ofthe non-capitalist or non- bourgeois classes in society. As we argue repeatedly, reduced inequality, even within this 'working class as a whole', requires job creation. Gelb and Webster appear to hold to a utilitarian view of social justice: the wage bill should be maximised regardless of distribution. Our approach is more of a Rawlsian one, in that priority should be given to improving the incomes of the poorest members of society. At the very least, the wage bill should not grow in ways that benefit the semi-privileged (ie people with jobs) at the expense of the poor. In the short-term, the government might commit more funds to public works programmes. These are not as politically difficult as reforming labour market policies, as COSATU has already accepted the principle that public works programmes can pay low wages as a way of targeting the very poor and destitute. The Working for Water Public Works Programme is widely regarded as a successful experiment, and could be expanded. Certainly, ifthe political will was there, resources could be transferred from unproductive spending (as is the case withmuch ofthe defence budget) to such programmes — potentially with a dramatic impact on employment. The problem is that unemployment in South Africa neither low nor temporary. In 1977, Chile employed 5.5 per cent of the labour force in emergency public works programmes, at wages of one third ofthe minimum wage (Cortazar 1997), but unemployment there was largely due to structural adjustment, and had declined to a very low level within a decade. In South Africa, public works programmes would probably have to continue for a much longer period. It is remarkable that the South African government has been so hesitant over the employment issue when it is such a pressing public priority. Opinion 10 Jeremy Seekings and Ntcoli Nattrass polls routinely show that job creation is the issue regarded by most voters as the most important problem facing the country. Yet the government seems almost incapable of action. Indeed, there is no single member of the cabinet withresponsibihty foremployment creation. Voters recognise this, dismissing government performance on the issue as unsatisfactory or very unsatisfactory. We hope that the democratic process will lead to heightened pressure on the government to respond to public opinion. (b) Education and Equal Opportunities Inequality in South Africa is closely related to education. On the one hand, tertiary education opens up the route to 'middle class' occupations. On the other, young men and women who leave school prior to matric are usually confined to a lifetime of low-paid employment or chronic or intermittent unemployment. How a student performs in school is, however, not simply a matter of individual ability. Rather, there is a clear relationship between parental class and children's educational attainment. A social democratic agenda must include the promotion of more equal opportunities. Under apartheid, massive investment in public education meant that the children of poor, mostly Afrikaans-speaking white workers did not end up with the same lack of skills as their parents (and were thus able to secure better-paid employment without explicit racial discrimination in the labour market). The challenge facing post-apartheid South Africa is how to promote better opportunities for children from poor backgrounds, regardless of race, but within die budgetary constraints set by the available tax base. Inequality in South Africa's schools is not primarily due to low enrolment rates. South Africa has high enrolment rates for all income deciles and regardless of gender (Lam 1999). Nor, or at least less and less, is it due to discrimination in spending: a larger share of public spending on education is spent on the poorest quintile than the richest quintile, whilst differences in spending per pupil are largely due to the fact that more of the children in the richer deciles are in secondary schools (which are more expensive to operate than primary schools). Rather, inequalities in education are due to differences in the quality of schooling,, as well as factors relating to the family background of pupils. The Department of Education has itself identified causes of poor performance: the poor educational background of parents, poor conditions of teaching and learning, inappropriate teaching and learning methods, lack of access to reading and other educational materials and libraries, poor school management, a lack of order and discipline among 20 Class, Distribution and Redistribution in Post-Apartheid SA both teachers and pupils (which often results in the loss oftime for teaching and learning) and the low morale ofprincipals and teachers (RSA 2000:40- 44). It is important to note that this list does not include pupil/teacher ratios, perse. Case andDeaton (1998), using the scanty statistics available, have shown that pupil/teacher ratios are not a major factor in educational outcomes. The Departments of Education and Finance, together with the Parliamentary Standing Committee for Public Accounts, have begun to pay attention to the problem ofgetting better' value-for-money' out of educational expenditure (see Seekings2001). School-children must achieve higher levels ofnumeracy, literacy and life-skills for the same level ofpublic expenditure. Research into this must pay careful attention to distributional issues, in that the goal of 'value-for-money' treats as valuable the achievement of substantially equal opportunity for children from whatever background. (c) Democracy-deepening asset redistribution Exploitation of waged employees by profit-seeking employers may not be ' the only source of-injustice in society, but it is clearly an important one. A social democrat needs to balance a concern with justice in the employment relationship with the need to keep firms operating so as to provide a measure of employment security for workers — whilst providing workers with a real say in the economic life and decisions of the firm. One way of doing this is to promote worker-ownership. The government might provide developmental support for worker-owned firms. Thus, for example, if a clothing firm decided that it could no longer compete profitably on international markets, the government should consider offering the workers (who face retrenchment) the option of taking over the firm and working for a share of profits rather than the wage. This would enable them to earn less than the going wage - as they would be drawing a share of profits rather than the old wage which had put the firm under pressure in the first place. Such an arrangement would at least keep people employed, albeit at lower earnings. However, there are reasons to believe that worker-owned firms can be more efficient than capitalist firms (Bowles and Gintis 1998). Thus, with government support (perhaps in the form of training, such as community service for MBA graduates), it is possible that worker-owners could end up earning more than they had as workers. But whatever the outcome, such creative solutions should be considered. The idea of promoting worker-ownership in manufacturing and services is hardly on the public agenda at present, and would probably provoke much 21 Jeremy Sseldngs and Nicoli Nattrass knee-jerk opposition. Yet there is widespread public support for worker ownership in another sector ofthe economy: agriculture. Economic arguments for land reform typically combine concerns with efficiency and distribution: not only will income be more equitably distributed, in mat there is no division between profits to a few and wages to the many, but also there will be more resources for the poor because production on small farms will be more labour-intensive than production on large farms. Implicit in this argument are the views that small farms will save on supervision costs (vis Bowles and Gintis) or will work at lower marginal returns than the prevailing wage for agricultural labour. There are, of course, other arguments for land reform: it provides security oftenure in accommodation, perhaps has social or cultural significance, and certainly has symbolic political importance ('Mayibuye i Africa'!). But government grants for land reform are certainly motivated in part by economic considerations. Since 1994 the government has provided subsidies for land reform, to enable African small farmers to buy land at market prices from white landowners. The government initially itself the target of transferring about 30 million hectares, or 30 per cent of the country's medium to high-quality farmland, to 600 000 households. To achieve this it provided a means-tested subsidy equal in value to the subsidy available for housing. Small farmers wbuld have to pool their subsidies and whatever other capital they could lay their hands on, and buy a farm from a landowner. It was not long before the Minister of Land Affairs decided that the plan was neither fiscally nor administratively viable. After three years, only 200 000 hectares had been transferred, to about 20 000 poor households (Deininger and May 2000:8). Deininger and May, in their studjrof the land reform experiment, concluded that, with minor reforms, it could make a major, contribution to reducing poverty in rural areas. But in 2000, the new Minister of Land Affairs and Agriculture announced a major shift in policy. The emphasis would henceforth be on commercial farmers, not small farmers or peasants. Government subsidies would be much larger for African farmers investing other funds of their own. In other words, the larger subsidies would be given to farmers with funds of their own already. The maximum grant would be a whopping R100000. Given that poverty is concentrated in rural areas, land reform needs to be examined carefully. If output can be increased if small farms practice more labour-intensive production, then land reform has the potential to promote growth and equity at the same time. Unless it takes place on a large scale, 22 Class, Distribution and Redistribution in Post-Apartheid SA however, it is unlikely to be a source of significant shifts in distribution. There is less poverty in towns and cities. But promoting worker ownership of firms in manufacturing and services entails a similar transfer of capital to workers as land reform, except that the capital involved is a factory or machinery rather than land. If land reform policy is reoriented toward the semi-privileged, ie prospective medium-sized farmers who can raise substantial funds themselves, why should the government not provide financial assistance to worker-owned firms? Why, we might also ask, do union investment companies not invest more in worker-owned enterprises? (d) Welfare Reform The final pillar in a social democratic agenda must be welfare reform. Public welfare is essential for the mitigation of acute poverty, and is especially just in a society where people are willing to work but cannot find work to do. The post-apartheid state inherited a welfare system that was remarkably redistributive in some respects (notably the old-age pension) but profoundly limited in others (most notably the absence of any support for most of the unemployed). Under Mandela's presidency the only substantial change to the welfare system was regressive, in that the value of public support for poor single parents was slashed (ostensibly to allow for higher take-up rates among the poor) and the real value of the old age pension declined. Soon after coming to office, however, the Mbeki government took the bold step of appointing the 'Taylor' Committee of Inquiry to propose reforms to achieve a comprehensive social security system. The Committee was charged with examining not just public welfare systems including unemployment insurance), but also public health and the integration of public and private welfare and health systems. One of the ideas prioritised by the Committee was the introduction of a basic income grant (BIG). The introduction of a BIG would make South Africa's welfare regime more coherent and appropriate for a middle-income labour-surplus economy. It could also help address the distributional dilemma ofthe HPN growth path in that those who gain (employed income- earners) would be taxed extra to pay for the grant. This was the implicit' social contract' behind the Scandinavian model - but it only held together while growth was rapid and unemployment relatively low (Nattrass 1999). But whether there is sufficient political will in South Africa to shoulder the additional tax burden remains to be seen. Samson et al argue that there is room for increased taxation because South Africa's average tax rate is below that of other countries at similar levels of 23 Jeremy Seekings and Nlcoli Nattrass development (2000:17). The 'People's Budget' (2002) (supported by COSATU, the South African Council of Churches and the South African NGO Coalition) proposes that part of the needed revenue could be raised through a 'solidarity levy' in the form of a 17.5 per cent surcharge on income tax for the top two quintiles - and the rest in the form of increased taxation of 'the high income group'. This is broadly in line with the COSATU 7th National Congress Resolution that the cost ofthe BIG must' fall on the rich'. This of course begs the question as to who comprises the rich. Is it simply the top decile, or is it the top 40 per cent of income earners? A discourse which refers confusingly to 'the rich' and 'the poor' is unlikely to contribute to the kind of social solidarity needed to sustain a BIG over the long-term. By arguing that 'the rich' should absorb all the costs of the BIG, the discourse of the People's Budget serves to reinforce class antagonisms rather than ameliorate them. As it is, many COSATU workers could find themselves paying the tax (even if only the proposed 'solidarity levy'), and could resist it because they had expected it to be levied on the 'rich' - a class they do not automatically associate themselves as being part of. In its 1998 submission to the Jobs Summit COSATU (1998) suggested that a BIG be financed in part by those earning over R3,000 a month paying back the amount they receive as tax, and those earning over R5,000 per month paying double the amount back. A high proportion of COSATU workers (particularly those in the government sector such as teachers and health workers) would fall into the R5,000 per month bracket - and in this regard, the proposal has clear social democratic aspects to it. However, by 2000, COSATU's position had hardened into a discourse that talked only of taxing 'the rich' rather than taxing income-earners. In fact, small increases in taxation on union members might be offset by reduced transfers from them to the poor through remittances, if the introduction of a basic income grant has any crowding out effects on remittances. If the reform of the public welfare system leads to a contraction of the private welfare system, it might be in the interests of trade unionists to tolerate small increases in taxation (see further Nattrass and Seekings 2002). The proposal for a BIG highlights a central policy dilemma in South Africa today: how to provide basic income support in a middle-income labour surplus society. Given that wage earners are protected relative to the unemployed, the moral and political quidpro quo is higher taxation in order to finance a BIG for the unemployed. The attraction is that the deal does not 24 Class, Distribution and Redistribution in Post-Aparttieid SA entail any erosion of labour standards or hard-won rights. The downside is higher taxation, and there is little indication that the organised working class is any more prepared to countenance much higher taxation than it is to agree to greater labour-market flexibility. Affirmative action: a note A social democratic policy agenda can accommodate affirmative action, as long as this is accurately organised along the lines of disadvantage in society. Affirmative action should preferably take the form of an investment in public education which ensures that all children face equal opportunities, regardless of background. Insofar as there are already many people in the labour market who did not enjoy equality of opportunity, there are strong arguments for affirmative action in the labour market also. But affirmative action must be organised according to an appropriate-measure of disadvantage. Despite the dramatic shifts in the determinants of inequality - and the increasing salience of class - inequality is often reduced to racial inequality in the minds ofmany South Africans. As we have noted elsewhere, President Mbeki's discourse of 'two nations' treats race as an accurate measure of disadvantage (Nattrass and Seekings.2001). The one nation is 'black and poor', the other 'white and relatively prosperous'. In terms of this world- view, affirmative action to address the legacy of racial discrimination is synonymous with reducing inequality. We have presented a range of evidence to the effect that inequality is increasingly a function of class, rather than race. During the apartheid era, there was a clear overlap between race and class. However, as Africans moved up the occupational ladder, and as the apartheid system frayed at the edges and then disintegrated, the contribution of within-group inequalitytototal inequality rose significantly. In post-apartheid South Africa, inequality is driven by two income gaps: between an increasingly multi-racial middle class and the rest, and between the African urban industrial working class and the African unemployed and marginalised poor. Race is thus an ever less adequate proxy for disadvantage. Affirmative action which simply ensures that better off African people move into posts previously occupied by white people might reduce inter-racial inequality but make no difference to overall inequality, as intra-racial inequality grows. This is what has happened, primarily as a result of the removal of discriminatory state restrictions rather than affirmative interventions, over 25 Jeremy Seekings and Nicoli Nattrass the past thirty years. If there is to be some redress of persisting inequalities of opportunity, then the marker of inequality shouldbe class rather than race. Because class is not an ascriptive characteristic, it has the additional advantage over race that it is unlikely to lead to the kinds of division often generated by race-based affirmative action (see also Ramphele 1996, Adam 1997). Conclusion The imagery of a divided nation is important in highlighting the extreme level of inequality in post-apartheid South Africa. It is always important to bear in mind the enduring effects of the racialisation of society under apartheid. But Mbeki's portrayal of 'two nations' is an inadequate analysis of South African society today. South African inequality is not simply or even primarily inter-racial. Declining inter-racial inequality has not reduced overall inequality, and will not do so in future, because the factors that drive inequality have become increasingly significant at the intra-racial level. In a society that has become dependent on wages and salaries, a reduction in inequality requires a more egalitarian and effective educational system, broader access to employment (through job creation) and reforms to the welfare system. Insofar as South Africa comprises a divided nation, it is more accurate to see it in terms of three broad social groups not two racially- de fined nations: an increasingly multi-racial upper class, comprising notjust high profile corporate figures but much more broadly the professional, managerial and business classes; a 'middle' group ofmostly urban, employed workers; and a marginalised class of outsiders, comprising many of the unemployed as well as workers in agricultural and domestic employment. Racial differences in income remain large (due mainly to differences in education and the labour market). But while changing the complexion of the rich is an important aspect of social transformation in post-apartheid South Africa, it should not be conflated with an egalitarian strategy. A strategy with redistribution at its heart would have to concentrate far more on creating j obs for the currently unemployed, and on redistributing income through the budget from rich to poor - perhaps through the introduction of a basic income grant. References Adam, K (1997) 'The politics of redress: South African style affirmative action', Journal of Modern African Studies 35(2). Class, Distribution and Redistribution in Post-Apartheid SA Adler, G and G O'Sullivan (1996) 'Rounding up the usual suspects: recycling the labour aristocracy thesis', in J Baskin (ed) Against the Current: labour and economic policy in South Africa. Johannesburg: Ravan Press. Adler, G and E Webster (1999) 'Towards a class compromise in South Africa's "Double Transition": bargainedliberalization and the consolidation of democracy', Politics and Society 27(3). (2000) 'Consolidating democracy in a liberalizing world: trade unions and democratization in South Africa', in G Adler and E Webster (eds) Trade Unions and Democratization in South Africa, 1985-1997. New York: St Martin's Press. Baskin, J (2000) 'Labour in South Africa's transition to democracy: concertation in a third world setting', in G Adler and E Webster (eds) Trade Unions and Democratization in South Africa, 1985-1997. New York: St Martin's Press. BMR (1999) 'Activities, lifestyles and status products of the newly-emerging middle class in Gauteng', Research Report 262, Bureau for Market Research, UNISA, Pretoria. Bhorat, H and J Hodge (1999) 'Decomposing shifts in labour demand in South Africa', South African Journal of Economics 67(3). Bond, P (2000) Elite Transition:from apartheid to neo-liberalism in South Africa. London: Pluto Press. Bowles, S and H Gintis (1998) 'Efficient redistribution: new rules for markets, states and communities', in S Bowles and H Gintis (eds) RecastingEgalitarianism. London and New York: Verso. Budlender, D (2000) 'Human development', in J May (ed) Poverty and Inequality in South Africa. Cape Town: David Philip. BusinessMap (1999) Black Economic Empowerment 1999: a moving experience. Johannesburg: BusinessMap SA. Carter, M and J May (1999) 'Poverty, livelihood and class in rural South Africa', World Development 27(1). Case, A and A Deaton (1999) 'School inputs and educational outcomes in South Africa', Quarterky Journal of Economics 458. Charney, C (1995) 'Voices of a new democracy: African expectations in the new South Africa', Research Report 38, Centre for Policy Studies, Johannesburg. Cichello, P, G Fields and M Leibbrandt (2000) 'Which African workers have gotten ahead and by howmuch?: the story of Kwa-ZuluNatal, South Africa: 1993-98'. Unpublished report, School of Economics, University of.Cape Town. Cortaza, R (1997) 'Chile: the evolution and reform of the labour market', in S Edwards andN Lustig (eds) Labor Markets in Latin America: combining social protection with market flexibility. Washington, DC: Brookings Institution Press. 27 Jeremy Seekings and Nicoli Nattrass COSATU (1998) Submission to the Jobs Summit. Available on www.polity.org.za/ govdocs/surnmit/jobsuinmit.html. Crankshaw, O (1997) Race, Class and the Changing Division of Labour under Apartheid. London: Routledge. Deininger, K and J May (2000) 'Is there scope for growth with equity? The case of land reform in South Africa'. Working Paper no. 29, Centre for Social and Development Studies, University of Natal, Durban. Edwards, L (2000) 'Globalisation and the skill bias of occupational employment in South Africa'. Unpublished paper, School of Economics, University of Cape Town. Gelb, S and E Webster (1996) 'Jobs and equity, the social democratic challenge', South African Labour Bulletin 20(3). Habib, A and V Padayachee (2000) 'Economic policy and power relations in South Africa's transition to democracy', World Development 28(2). Heintz, J and S Bowles (1996) 'Subsidising employment: wage subsidies and job creation', in J Baskin (ed) Against the Current: labour and economic policy in South Africa. Johannesburg: Ravan. Kenny.B andE Webster(1998) 'Eroding the core: flexibility and the resegmentation of the South African labour market', Critical Sociology 24(3). Lam, D (1999) Generating Extreme Inequality: schooling, earnings and inter- generational transmission of human capital in South Africa and Brazil. Research Report # 99-439, Population Studies Center at the Institute for Social Research, University of Michigan. Maiais,H(1998)SouthAfrica, Limits to Change: thepolitical economy of transition. Cape Town: University of Cape Town Press and London: Zed Books. ( 2001) South Africa, Limits to Change: thepolitical economy of transition. Second edition. Cape Town: University of Cape Town Press and London: Zed Books. McGrath, M, J Janisch and C Homer (1997) 'Redistribution through the fiscal system in the South African economy'. Paper presented to the Economics Society of South Africa Conference, Potchefstroom. Michie, J and V Padayachee (1998) "Three years after apartheid: growth, employment and redistribution', Cambridge Journal of Economics 22(5). Nattrass, N (1999) 'Globalisation and social accords: a comparative analysis of Sweden, Australia and South Africa', Labour, Capital and Society 32(2). (2000a) 'Inequality, unemployment and wage-setting institutions in South Africa', Studies in Economics and Econometrics 24(3). (2000b) "The debate about unemployment in the 1990s', Studies in Economics and Econometrics 24(3). 28 Class, Distribution and Redistribution in Post-Apartheid SA (2001) 'High productivity now: a critique of South Africa's growth strategy', Transformation 45. Nattrass, N and J Seekings (1997) 'Citizenship and welfare in South Africa: deracialisation and inequality in a labour-surplus economy', Canadian Journal of African Studies 31(3). (1998a) 'Democratic institutions and development in post-apartheid South Africa', in M Robinson and G White (eds) The Democratic Developmental State: political and institutional design, Oxford Studies in Democratisation. Oxford: Oxford University Press. (1998b) 'Growth, democracy and expectations in South Africa', in I Abedian and M Biggs (eds) Economic Globalisation and Fiscal Policy. Cape Town: Oxford University Press. (1998c) 'Changing patterns of inequality in the South African labour market', in LPetersson(ed) Post-ApartheidSouthern Africa. London: Routledge. (2001 a) 'Democracy and distribution in highly unequal economies: the case of South Africa', Journal of Modern African Studies 39(3). (2001b) 'Two nations? Race and economic inequality in South Africa today', Daedalus 130(1). (2002) 'A basic income grant in South Africa?'. Paper presented at the First Congress of the United States Basic Income Guarantee Network, New York (March). - (forthcoming) FromRace to Class: inequality, unemployment and the social structure in South Africa. Cape Town: David Philip. People's Budget (2002) 'The people's budget'. Available on www.cosatu.org.xa/ docs/2001 /peobudO 1 .htm. Posel, D (1991) The Making of Apartheid, 1948-1961. Oxford: Clarendon Press. RSA (2000a) Budget Review 2000. Department of Finance. (2000b) Educationfor All: The South African assessment report. Department of Education, Pretoria (March). Ramphele, M (1996) 'Treading the thorny path to equity', in W James et al (eds) Now that We are Free: coloured communities in a democratic South Africa. Cape Town: IDASA. Randall, DJ (1996)' Prospects for the development ofa black business class in South Africa', Journal of Modern African Studies 34(4). Samson, M, O Babson, K MacQuene, I Van Niekerk and R van Niekerk (2000) "The macroeconomic implications of poverty-reducing income transfers'. Paper • presented to the conference: Towards a Sustainable and Comprehensive Social Security System, hosted by the Institute for Social Development, University of the Western Cape. 29 Jeremy Seekings and Nicoli Nattrass Seekings, J (2000) 'Visions of society: peasants workers and the unemployed in a changing South Africa', Studies in Economics and Econometrics 24(3). (2001) 'Making an informed investment: improving the value of public expenditure in primary and secondary schooling in South Africa'. Report for the Parliament of South Africa, commissioned by the Standing Committee on Public Accounts (December). Simbi, T and M Aliber (2000)' The agricultural employment crisis in South Africa'. Paper presented at the TIPS Policy Forum, Muldersdrift, September. Simkins, C and I Woolard (2000) 'Analysis of the burden of taxes in the South African economy for the years 1995 and 1997'. Report for the Department of Finance. South African Health Review (1997). Standing, G, J Sender and J Weeks (1996) The South African Challenge: restructuring the South African labour market. Geneva: International Labour Organisation. Theron, J (2000) 'Responding to externalisation: Part 1', South African Labour Bulletin 24(6). Van derBerg, S (2000b) 'An analysis of 1997 school data flowing from work on the fiscal incidence of social spending in South Africa'. Report for the Department ofFinance. (2001) 'An analysis of fiscal incidence of social spending in South Africa, ' 1993-97', Social Dynamics 27( 1). Webster, E (2001) 'The alliance under stress: governing in a globalizing world', in R Southall (ed) Opposition and Democracy in South Africa. London: Frank Cass. Whiteford, A and D van Seventer (2000) 'Understanding contemporary household inequality in South Africa', Studies in Economics and Econometrics 24(3). 30