LABOUR TRENDS OBSERVED IN SOUTH AFRICA:

Transcription

1 DIFID-WB Collaboration on Knowledge and Skills in the New Economy LABOUR TRENDS OBSERVED IN SOUTH AFRICA: A context paper prepared for the World Bank by Servaas van der Berg & Megan Louw University of Stellenbosch August 2004 INTRODUCTION The reference paper shows that the relative demand, supply and wages for workers with higher levels of educational attainment are rising. There has been a sharp increase in the supply of workers with secondary education relative to those with primary education, and a small increase in the supply of workers with tertiary education relative to those with secondary education. The relative wages for workers with secondary or higher education have risen steeply despite the increase in the supply of workers with greater educational attainment. Sànchez-Pàramo and Schady (2003: 4) comment that increases in the relative wages of skilled workers that occur simultaneously with increases in their relative supply provide strong evidence of an increase in the demand for skilled workers. Indeed, the relative demand for workers with higher levels of education is rising over time. This context paper aims to shed some light on the observed labour trends through reviewing changes in South Africa s macroeconomic structure, education system and labour market over the period Understanding the context within which the South African labour market operates is crucial for an overview of labour market trends. A number of features of the South African labour market need to be understood: The pervasive legacy of apartheid conditions responses of unions and government to labour market outcomes and inequalities. The virtual absence of a peasantry and an unusually small informal sector limit survival options, thus large-scale unemployment is more visible than in many comparable economies. The magnitude and growth of unemployment over the entire educational spectrum reduce the strength of the conclusions one can draw about relative wage trends, since wages respond more sharply to institutional factors than to relative scarcities in a labour market that does not clear. Finally, the heterogeneity of measured educational attainment over time, space and especially race groups makes it difficult to draw hard conclusions pertaining to the productive characteristics of the workforce at various educational levels, since these levels mask large variation in the quality of education obtained by workers. SOUTH AFRICA S MACROECONOMIC STRUCTURE Historically, the South African economy has been driven by agriculture and mining, although over the past sixty years the contributions made by the manufacturing and service sectors to GDP have increased substantially. In 2002, 64 percent of the economy was accounted for by the tertiary sector, 24 percent by the secondary sector and only 12 percent by the primary sector (South African Reserve Bank 2003). This primary sector is largely composed of mining, while agriculture is exceedingly small for a developing country, making a

2 2 contribution of only 4 percent to the economy. This change in the structure of the economy has accelerated since 1994, with the reintegration of the country into world capital and product markets following political transition. South Africa has engaged in substantial trade liberalisation during the past decade since becoming a signatory to the General Agreement on Tariffs and Trade in 1994 and thus formally rejoining the international trading community. In fact, the current guiding macroeconomic strategy documented in GEAR (Growth, Employment and Redistribution) revolves around export-led growth and as such has relied upon a reduction in the country s persistent anti-export bias (Tsikata 1999: 3). As a result of South Africa s policy stance and the accompanying reform measures, total trade (exports plus imports) as a share of GDP had risen to 64 percent by 2002 (South African Reserve Bank 2003), compared to the average total trade to GDP ratio for non-oil middle-income countries of around 46 percent in 1999 (Tsikata 1999: 8). An increasingly high level of involvement in international trade has led to significant changes in the composition of the South African economy, prompting Fedderke et al (2001) to suggest that the early 1990s signal a structural break. Progress towards a knowledge economy has been observed, with more highly skilled sectors (particularly those producing technologyintensive manufactures such as machinery and metal products: Tsikata 1999: 27) making substantial gains in relative importance. Further, manufacturing exports have expanded rapidly and the aggregate export base has diversified significantly as a result of changes in relative prices and incentives facing exporters, as well as a post-sanctions effect (Tsikata 1999: 19). On the whole, there has been a continuing shift towards capital-intensive exports. The losers from trade liberalisation have been labour-intensive sectors such as agriculture and mining, as well as unskilled labour-intensive manufactures including textiles (Tskikata 1999, Fedderke et al 2001). Overall, then, South Africa s share of exports that use unskilled labour is declining rapidly, while the share of exports that use more skilled labour and technology is increasing in relative importance. From a comparative advantage perspective, these trade patterns are counterintuitive, given the country s factor endowments South Africa has abundant supplies of unskilled labour (Tsikata 1999: 26). One would expect a trading country to specialise in the production of goods that use its abundant factor intensively, causing a rise in the price for the abundant factor as a result of the greater demand for it. However, in South Africa policies directed at bringing the cost of capital down to artificially low levels as well as political and labour unrest prior to political transition in 1994 skewed production methods in favour of capital intensity. This trend towards capital intensity has continued since the political transition, with the opening up to international trade of the mid 1990s, global technology trends, and the relatively high cost (and low productivity) of unskilled labour in South Africa resulting in production specialisation patterns very different from those that theory predicts. Reviewing the development of the macroeconomic structure has been the first step in understanding the observed labour trends; attention turns now to a more focused analysis of how changes in the education system over the past decade have affected the supply of workers with different levels of education. SOUTH AFRICA S EDUCATION SYSTEM Under the apartheid system, separate schooling systems were set up for members of different race groups. Black schools were granted significantly smaller but highly variable teaching resource allocations compared with those given white schools. As a result, schools became

3 3 highly stratified in terms of the quality of education provided. Following transition to democracy, the different schooling systems were re-unified, pupil-teacher ratios were standardised across schools (they currently stand at 32) and teacher salaries were equalised across racial lines and significantly raised (in response to teacher union pressure). While public spending on education has increased dramatically to a level of around 23.5 percent of the 2002/03 budget (South Africa, National Treasury 2004: 208), large fiscal resource shifts have not been accompanied by proportionate real resource shifts, since much of the funding has gone towards equalising teacher salaries. The result has been that while educational expenditure nears its limit, little has changed during the past decade in terms of the quality of schooling that the majority of learners receive. Formerly whites only schools, now deracialised, continue to provide good quality education supported by access to the best teacher resources in the public school system. Most formerly black schools and these comprise the bulk of schools in the new system still lack more qualified teachers, and the administration and management skills with which to efficiently manage the resources they have available to them. As a result these schools continue to under-perform. Indeed, widespread ongoing problems with quality have resulted in very unfavourable rankings in international education comparisons (for example the World Competitiveness Yearbook referred to in Poswell 2002: 19, or the 1995 MLA study in which the numeracy performance of South African students was the worst of twelve African countries surveyed). Turning to issues regarding the quantity of education provided, government has fared well in achieving almost universal enrolment up to age 15 (i.e. up to grade 9). This has been the result of rapidly expanding primary education (the first seven years of school) during the 1960s, and secondary education (a further five years) during the 1970s and 1980s. At higher secondary grades, there are high drop-out rates linked to the weak quality of education offered at many of the poorer schools, although a large number of learners remained in school into their twenties, possibly due to their negative perceptions of their prospects in the labour market. As a result, the Department of Education has placed restrictions on over-age children remaining in the school system. In spite of the problems plaguing the schooling system, secondary school completion rates are rising. At the end of Grade 12, students write a matriculation or senior certificate examination, largely assessed provincially but with standards set nationally. In 1990, learners matriculated; by 2003, this figure had grown to But to enter university, an even better performance in the matric examination, plus appropriate subject choice, are usually required. The growth in matriculation exemptions or endorsements (needed to qualify to apply to universities) has been slower over this period, from a level of to a level of The impact of the expansion of primary and secondary schooling can be seen in the figure below, which shows the distribution of education by birth cohort.

4 4 Education level completed by birth cohort as at Census % 90% 80% 70% 60% 50% 40% 30% 20% 10% Tertiary Complete secondary Primary & incomplete secondary None & Incomplete primary 0% Year of birth With respect to the tertiary education system, growth in numbers of university students has been relatively slow over the past decade, averaging 2.3% per annum over This has reflected the modest growth in university exemptions, poorer students lack of access to university funding, demographic changes and perceptions that a university education obtained in one of the less technical fields does not necessarily improve an individual s job prospects. In a recent study, Bhorat (2003b) notes the beginning of a graduate unemployment problem. In parallel with the deracialisation of the schooling system, the tertiary education system has experienced significant transformation over the past decade, with restructuring along nonracial lines involving inter alia the amalgamation of some tertiary institutions. There has also been a large uptake of black students in historically white universities. However, as in the schooling system, considerable variation still exists in the quality of tertiary education offered at historically black and historically white universities (Bhorat 2000a: 8). The 2001 census shows that 42.9 percent of the 25 year old age group (largely new or otherwise recent entrants to the labour force) had completed at least matric, versus only 13.6 percent of the 65 year old age group which are largely those exiting the labour market. At the midpoint between these two birth cohorts, the percentage of 45 year olds completing matric was 21.2 percent, showing that progress in expanding schooling has accelerated in the last two decades. As a result of the changing education structure, the proportion of the prime labour force age group (20-64 years) with complete secondary or tertiary education increased from 26 to 30 percent between 1995 and Based on sample surveys, the table below shows how the educational profile of the South African labour force has changed over the period :

5 5 Educational profile of South Africa: Individuals aged 26 years and over No schooling 15.78% 12.64% Incomplete primary education 19.11% 18.65% Primary complete education 7.30% 7.79% Incomplete secondary education 32.32% 31.96% Secondary complete education 15.97% 18.51% Tertiary education 9.52% 10.45% % % Source: own calculations based on October Household Survey 1995 and Labour Force Survey September 2002 Aside from the progress made in formal education, government also launched a National Skills Development Strategy aimed at improving the workplace skills base from which employers can draw. Employers are required to pay 1% of their payroll as a skills development levy, part of which can be reclaimed to fund personnel training. Allowance is also made for the training of targeted vulnerable groups, and for learnership agreements in terms of which new labour market entrants receive on-the-job training from employers in exchange for the provision of employer tax benefits. Implementation of the strategy would be expected to boost economy-wide productivity and growth potential, although progress to date has been hampered by insufficient capacity and bureaucratic inefficiency (Lundall 2003). SOUTH AFRICA S LABOUR MARKET Apart from education, other factors affecting the labour supply derive from institutional structures including unions, bargaining structures and labour legislation, and from crossborder migration. Following political transition, a number of labour laws were passed to protect workers and to eliminate the racial character of the labour market. During the apartheid years, racial discrimination played a large role in determining patterns of employment and wage differentials, so that whites filled most skilled (and better paying) positions, while blacks filled many semiskilled or unskilled (typically low paying) positions. Consequently, the new government introduced legislation aimed primarily at protecting workers at the middle and lower end of the skill spectrum (Labour Relations Act, 1996 and Basic Conditions of Employment Act, 1997), as well as affirmative action legislation promoting blacks in the workplace (Employment Equity Act, 1998 and the more recent sectoral charters). While these measures support a move towards greater equity in the labour market, there are concerns that they have made the labour market less flexible through making dismissal of employees more difficult, giving trade unions more power, and so on (Fallon & Lucas 1998). There are also fears that affirmative action policies may lead to new forms of labour discrimination, this time in favour of the formerly disadvantaged groups, and that once again the country may for political reasons not fully utilise its labour skills. With respect to trade unions, the membership rate in South Africa is around 34%, substantially higher than the developing country average of 18% (Poswell 2002: 15). As a result of apartheid policy less skilled workers were paid artificially low wages, which unions have worked to raise. Employers may have responded to the introduction of the recent labour legislation or the increased economic bargaining strength of trade unions by cutting back on new hires of

6 6 unskilled or semiskilled workers or by substituting capital for labour. Finally, international migration has increased the supply of unskilled labour as well as some skilled labour from neighbouring African countries, but an ongoing brain drain to developed countries is decreasing the supply of skilled labour. Next, attention turns to the implications of the evolution of the South African economy for the demand for labour. The major employer sectors in South Africa are the agricultural, government, gold mining and service sectors (Fedderke et al 2001: 5). Bhorat (2003a) shows that employment growth over was particularly strong in the financial and business services sector, followed by construction and internal trade. It should be noted that this is based on household surveys, however, and that their quality and comparability (e.g. in terms of definitions of occupations) are not always certain. Negative growth in employment was reported for community, social and personal services (constituted mainly by government the largest employer in the South African economy), mining, utilities and agriculture. The reduction in employment in the community, social and personal services sector is due to extensive public sector restructuring aimed at reducing inefficiencies and the burden of the public wage bill. Employment losses as a result of rationalisation and restructuring were greatest for two categories of less skilled workers unskilled or elementary workers and machine operators although there were also losses in employment for skilled workers 1 with tertiary education (Poswell 2002). At the economy-wide level, occupational groups experiencing drops in the demand for their labour were technicians, clerks, machine operators and unskilled workers (Bhorat 2003a). Substantial increases in labour demand were recorded for professionals and managers. However, it should be borne in mind that the largest part of the workforce is still the unskilled component (Fedderke et al 2001: 15). Analysing shifts in relative labour demand, Bhorat (2003a) finds that there were relative increases for managers as well as semiskilled workers such as clerks and sales personnel. The rising demand for semiskilled workers was driven by the demand for labour in the growing sectors of construction, internal trade and finance. The relative demand for skilled workers other than managers (i.e. professionals and technicians) declined, although this seemed to be entirely due to public sector restructuring; substantial increases in the demand for both categories of workers were observed when the community, social and personal service sector was removed. With respect to unskilled workers, relative demand declined over the period Overall, Bhorat (2003a: 13) notes that the greatest increases in the relative demand for labour in South Africa over the period were for semiskilled and skilled workers. This accords with the finding in the present paper that the relative demand for more skilled workers (individuals with secondary and tertiary education) is rising. Which factors have been driving the increasing demand for skilled labour? The Hecksher- Ohlin and Stolper-Samuelson theories predictions for production specialisation and factor prices would lead one to expect that an increase in the relative size of the skill-intensive sector may cause a national skills upgrade, in response to the greater demand and pay for skilled labour. One would expect the skills upgrade to result in within-sector substitution towards relatively inexpensive unskilled labour, unless skill-biased technological change was present, driving the increased relative wages and employment shares of skilled workers within sectors (Sànchez-Pàramo & Schady 2003: 14). Trade may act as an important mechanism for the transmission of such skill-biased technological change, as domestic firms faced with the 1 Here, skilled workers refer to managers, professionals and technicians, while semiskilled workers include clerks, service and sales workers, skilled agricultural workers, craft workers and machine operators.

7 7 threat of international competition upgrade their methods of production in order to remain profitable. In a recent study employing the Katz and Murphy (1992) methodology for decomposing changes in labour demand, Bhorat (2003a: 15) finds that within-sector rather than between-sector forces have been driving changes in employment in South Africa since This suggests that skill-biased technological change, changing preferences for production factor mixes, and other firm- or sector-specific factors are the major forces behind observed labour market trends. It also implies that structural change within the economy (i.e. change in the relative economic importance of sectors) has been far less important for labour market trends than the technological changes necessitated by international re-integration and modernisation. Bringing what we know about labour demand and supply in South Africa together, Bhorat (2003a) notes that there is a continuing skills mismatch in the labour market, reflected in skills-biased labour demand shifts. Lower skilled labour is being shed, while increases in labour demand are mostly concentrated at the upper end of the skills continuum. Further, while the demand for skilled labour is rising, the demand is for very specific types of skills provided by skilled workers, which may be the reason why there is rising unemployment even amongst graduates (Bhorat 2003b). Unemployment for all young workers is substantially higher than it is for the adult population (in line with the general international experience): Mlatsheni and Rospabe (2002: 3) estimate that 58 percent of year olds were unemployed in 1999, compared with only 26 percent of the adult population (i.e. those aged 31-65). Bhorat (2003b: 19) finds some evidence that youth unemployment is worsening, since the age cohort accounted for 30 percent of the unemployed in 1995, compared with 34 percent in The high rate of youth unemployment can mostly be attributed to a lack of work experience, and despite rising youth unemployment even amongst the more educated, it is worst amongst youth with incomplete secondary schooling or less education (Mlatsheni & Rospabe 2002). Furthermore, unemployment at the aggregate level continues to rise, as growth in the supply of labour has been larger than the growth in labour demand since 1995, leading to a worsening in South Africa s extremely serious unemployment problem: it is said to be the only medium-sized country in which there are more unemployed individuals than informal sector workers (Fallon & Lucas 1998: 1). Employment in mining, post and telecommunications, and community, social and personal services declined between 1995 and 2002, contributing towards the decline in the share of unskilled workers in the labour force from 31 percent in 1995 to 27 percent in 2002 (Bhorat 2003b: 11-12). Features of the economy that exacerbate the unemployment problem are the almost complete absence of a peasantry and the small contribution of agriculture to GDP, meaning that South Africa lacks two labour absorbing sectors (the informal sector and peasant agriculture) that often act to reduce visible unemployment in other countries. In 1995, the narrow unemployment 2 rate was 17.0 percent (October Household Survey 1995); by September 2002, it had increased to 30.5 percent. However, these figures do not include discouraged work seekers, and so understate the true extent of the problem in a country where unemployment is so widespread that many individuals drop out of the labour force after a period of searching for work. Broad unemployment (which includes discouraged workseekers) has risen from around 30 percent in 1995 (October Household Survey 1995) to 41.8 percent in September 2002 (South Africa, 2 This is effectively the same as the International Labour Office definition of unemployment (Fallon & Lucas 1998:5)

8 8 Statistics South Africa 2003). Unemployed individuals are more likely to be black and female, and to have low education levels (Bhorat 2003b: 14-15). The table below contrasts the proportions of individuals with each level of education who are employed or unemployed (by the broad definition) at the beginning and end of the period under study. Employment profile of South Africa: Employed Unemployed Employed Unemployed No schooling 8.17% 10.29% 19.64% 22.34% Incomplete primary education 15.96% 21.55% 24.09% 27.04% Primary complete education 6.51% 9.03% 6.40% 7.44% Incomplete secondary education 31.83% 38.61% 25.99% 28.68% Secondary complete education 22.86% 18.25% 15.07% 11.69% Tertiary education 14.67% 2.27% 8.81% 2.82% % % % % Note: unemployment is by the broad definition Source: Own calculations based on October Household Survey 1995 and Labour Force Survey September 2002 Note that the proportions of employed individuals at all levels of educational attainment are declining, and that there has been a particularly large relative drop in employment for tertiary educated workers (it should be noted that this includes very heterogeneous types of education, however). Official figures show that employment declines may not fully capture trends towards out-sourcing of work. Survey estimates that suggest positive job growth since 1995 probably reflect reality more accurately, although there can be no doubt that job creation has not kept pace with growth in the labour force which leads one to conclude that the labour absorptive capacity of the economy continues to be grossly inadequate. Another interesting finding with respect to tertiary educated workers is that there is no gender gap in the incidence of employment (in fact, slightly more females with tertiary education are employed than their male counterparts). This is in sharp contrast with the findings for the incidence of employment for workers with primary and secondary schooling, where men dominate by a significant margin. Bhorat (2003a) argues that the rapid recent employment growth in financial services and internal trade sectors seems to have benefited women more than men, and that growth in the demand for clerical and sales staff has been especially beneficial for female employment. Females are also more inclined to employment in teaching and nursing, which required tertiary education. Linking intra-sectoral employment changes to the educational attainment of workers, one would expect particularly the financial services sector to absorb skilled female workers due to the sector s high degree of skill intensity. Bhorat (2003a) argues that the gains to female workers as a result of job creation in highgrowth sectors outweighed losses in male employment due to the latter workers bearing the largest unemployment burden in declining sectors over , leading to higher growth in the net demand for female workers over this period. With respect to changes in relative employment by gender, Casale (2004: 11) finds that there were significantly larger increases in the number of employed women with degrees than similarly qualified men over the period Further, she observes that women with degrees benefited from average real

9 9 increases in earnings over the same period, while men did not 3. The feminisation of the labour force may be due to factors including increased educational attainment by women or declining material support from men (Poswell 2002: 8). On the demand side, employment equity legislation is likely to be contributing towards increases in female employment. Further, Bhorat (2003a: 16) suggests that changing production methods as a result of the microelectronics revolution as well as a growing preference for capital over labour have lowered the entry barriers to the labour market for women in South Africa. AN EARNINGS PROFILE OF SOUTH AFRICA Having considered the factors driving labour demand and supply, attention shifts towards identifying the major factors influencing earnings in South Africa. A substantial body of research on the topic has appeared since the mid 1990s. The international literature on the rate of return to schooling (see for example Psacharopoulos (1994)) generally reports diminishing wage returns to education, i.e. that the earnings curve is concave. However, recent research using South African data has arrived at results that are inconsistent with this finding (e.g. Hosking 2001; Fryer & Vencatachellum 2003). Using South African data from 1993 to 2000, Keswell and Poswell (2002) find that the relationship is in fact convex that there are increasing returns to education, with fairly low and flat marginal private returns to primary and secondary schooling that subsequently rise steeply with the acquisition of tertiary education. In a non-linear model with a cubic education term, the return to primary education is 1 percent, to incomplete secondary education it is 15%, to complete secondary education 32 percent and to tertiary education 54 percent (Keswell and Poswell 2002: 38) The finding that the earnings curve is convex is in accordance with what we know about the labour market: the supply of workers with only primary or to a lesser extent secondary education is large, although relatively low secondary school completion rates and particularly the poor quality of matriculation passes have put limits on the size of the pool of individuals who are eligible for tertiary education. In addition, the quality of education received by most black learners at primary or secondary school level is poor, and as a result educational attainment may overstate cognitive learning. It is possible that employers offer lower wages to black individuals with secondary schooling or less when taking this into account. Indeed, Case and Yogo (1999) find that the quality of schooling (proxied by the pupil-teacher ratio) received by black men under apartheid significantly affected the rate of return to their educational investment. Further, the bottleneck in learner numbers at upper secondary school level exists in the context of a skill-intensive economy, so that those with tertiary educations (largely whites) receive wage premia for both being in the smallest supply and in the greatest demand. Entry into university remains constrained both by funds and by the quality of school education (to obtain university endorsements, students usually need good matric results and an appopriate subject choice at school), thus many of the poor find it difficult to obtain tertiary qualifications. Apart from being influenced by differences in productive characteristics of workers, earnings in South Africa are also influenced by a range of non-productive characteristics. Fallon and Lucas (1998: 14) comment that wage differentials in South Africa are heavily influenced by factors that should have little bearing in a well-functioning labour market. The first of these factors is worker discrimination, of which the best publicised example in the South African 3 However, it should be borne in mind that within race groups, women receive lower returns to education than men do, regardless of their level of education (Casale 2004: 11).

10 10 context is race. The literature has often attributed the large racial earnings gap still observed in South Africa (white workers receive significantly higher wages than black workers) to employer discrimination, although Chamberlain (2001) highlights differences in the quality of education received by blacks and whites in explaining racial wage differentials. In his study based on data for 1995, Chamberlain (2001: 90) found that the differential in the rate of return to whites and blacks almost halved (from 42 percent to 24 percent) when education quality was controlled for. Hertz (2003) adds that ignoring the low and highly variable quality of education in black schools results in estimated rates of return to education for blacks that are biased upwards; after controlling for school quality, the marginal private return to year of schooling is only 5 percent, half of what it is before controlling for quality. However, education issues aside, it is likely that an element of race-based discrimination remains operative in the labour market. Strong gender discrimination has also been observed in South African earnings, although it does not seem to be as substantial as racial discrimination. Female workers earned 35 percent less than males in 1995, a gap that barely narrowed to 34 percent in 2001 (own calculations based on Casale 2004: 7). Once again, a large part of the gender differential in earnings may be due to factors other than overt discrimination, including women being more likely to work in low-paid domestic work or in informal sector self-employment (Casale 2004: 13). The observed small decline in the gender wage differential over the period under study is likely to be a result of an increase in the relative demand for female workers, as discussed earlier. Secondly, there are various barriers to mobility in South African labour markets that drive wage differentials, typically location of employment and whether a worker has informal or formal employment status (Fallon & Lucas 1998: 14). Segmentation in the labour market prevents participants from moving freely and thus competing with each other on the basis of their productive characteristics. In South Africa, research indicates that a high degree of segmentation exists between formal and informal sector workers (Hofmeyr 2001). Since the 1970s, the labour absorptive capacity of the South African economy has not been strong enough to match the growth in labour market participants. As a result, the informal sector has grown in relative importance. Over the period , surveys from Statistics South Africa (referred to in Devey et al 2003) have shown that the informal sector roughly doubled in size to around 1.9 million individuals 4. Work in the informal sector is generally insecure and poorly paid: using 1999 data for black male workers, Hofmeyr (2001) finds that informal sector workers with the characteristics of the average formal sector worker earn around 31 percent less than their formal sector counterparts. No trend can be identified though, as the author comments that the finding for 1999 cannot be contrasted with the 1995 results due to problems with the collection of data for informal sector workers in the earlier year. In addition to employment sector, location also matters for earnings. In contrast to the bulk of developing countries, South African labour is mostly urban: less than 40% of the workforce is found in rural areas (Fallon & Lucas 1998: 3). Fallon and Lucas (1998: 14) estimated that in 1993 black workers in large metropolitan areas earned 66% more than black workers in rural areas, while white workers in metropolitan areas earned 24% more than white workers in rural areas. A third factor driving wage differentials is variation in the characteristics of workers who have found employment in the formal sector, for example union membership. Recent studies by Rospabe (2001) and Bhorat and Leibbrandt (2001) find that the union wage premium is 4 However, Devey et al (2003: 45) argue that this data is variable and unreliable.

11 11 around 20 percent for black workers, but that there is no such unionisation premium accruing to white workers. Analysing data for black male workers, Hofmeyr (2001) argues that the union wage premium increased significantly between 1995 and 1999, largely as a result of the introduction of labour legislation discussed earlier and its implications for union power in wage negotiations. He estimates that an average non-unionised formal sector with the same characteristics as the average unionised formal sector worker earned 11.5 percent less than his unionised counterpart in 1995, and 20 percent less in On this basis, Hofmeyr (2001: 11) concludes that the degree of labour market segmentation is worsening: the unionised segment of the workforce has managed to successfully protect its earnings while the non-unionised segment has failed to do the same. IS THERE EVIDENCE OF SKILL INFLATION? Finally, we address the issue of skill inflation in the South African context. This provides a competing explanation for the observed changes in the skill profile of the labour force. If skills inflation were taking place, employers would be hiring more skilled workers to fill the positions previously occupied by other workers (bear in mind that South African unskilled labour is relatively expensive). A preference for higher skill levels would thus reflect employer preferences for efficiency and price gains in the face of the greater competitive pressures brought by international trade (Bhorat 2000b: 458). The table below reflects the changing skill profile of the manufacturing sector over the period under study, at the level of occupational grouping (skilled agricultural workers are excluded as there were very few of them working in manufacturing). The manufacturing sector contributed 19 percent of the economy in 2002 (own calculations based on South African Reserve Bank 2003), and provides a large share of South African exports. As can be observed, there is no clear evidence here of skills inflation, but that may in part be an artefact of the poor comparability of the two surveys, in terms of occupational structure. At this stage there is no hard evidence of such skills inflation, although it is quite likely that some examples of this are to be found in the labour market, particularly amongst young work seekers.

12 12 A profile of educational attainment by occupation within the manufacturing industry, Educational attainment Manager Professional Technician Clerk Sales Craft Operator Elementary No schooling 1.22% 0.00% 0.00% 0.19% 4.89% 5.67% 5.07% 9.70% Incomplete primary education 2.86% 0.00% 0.00% 0.51% 9.75% 9.80% 11.38% 21.34% Primary complete education 0.59% 0.00% 0.61% 1.00% 7.03% 7.40% 8.92% 14.40% Incomplete secondary education 16.36% 0.00% 14.60% 43.62% 45.11% 46.09% 51.89% 40.97% Secondary complete education 44.40% 1.87% 43.89% 43.28% 27.22% 25.64% 18.99% 11.97% Tertiary education 34.57% 98.13% 40.91% 11.40% 6.01% 5.41% 3.75% 1.62% % % % % % % % % 2002 Educational attainment Manager Professional Technician Clerk Sales Craft Operator Elementary No schooling 5.97% 8.51% 12.26% 10.49% 13.79% 19.12% 14.32% 16.78% Incomplete primary education 11.49% 6.42% 15.64% 14.22% 26.44% 24.88% 24.03% 23.15% Primary complete education 2.03% 0.68% 3.74% 3.62% 6.51% 7.69% 8.27% 8.93% Incomplete secondary education 22.34% 16.15% 33.11% 28.80% 35.34% 29.61% 33.77% 33.36% Secondary complete education 25.61% 18.73% 20.91% 32.84% 16.60% 13.78% 16.86% 15.27% Tertiary education 32.57% 49.51% 14.35% 10.03% 1.31% 4.92% 2.75% 2.51% % % % % % % % % Source: own calculations based on October Household Survey 1995 and Labour Force Survey September 2002 CONCLUSION This paper has shown that there is indeed strong evidence of a shift in labour demand towards more skilled labour in South Africa. In the South African context, higher skills usually require more than secondary education only, and the supply of such tertiary educated workers has not expanded very rapidly due to quality constraints at the secondary education level, amongst other things. However, in the case of secondary educated workers, this supply has increased very strongly, to the extent that there has been a sharp increase in unemployment amongst secondary educated workers 5. Given this large rise in unemployment amongst the secondary educated, their strong wage rises relative to unskilled or primary educated workers may not be so much the consequence of a rise in the demand for the former, as a decline in the demand for primary educated workers. Moreover, institutional factors in particular the role of trade unions may have played a more important role in driving wage movements than the relative scarcities of workers with different levels of education. BIBLIOGRAPHY ANDERSON, K.G Causes and consequences of schooling outcomes in South Africa: Evidence from survey data. Research Report Ann Arbor: Population Studies Center, University of Michigan. October. BARKER, F.S On South African labour policies. The South African Journal of Economics 67(1): March. 5 And indeed also amongst some tertiary educated workers, but this may be largely linked to specific training and the heterogeneity of education quality and content amongst the tertiary educated.

13 BHORAT, H. 2000a. Wage premia and wage differentials in the South African labour market. DPRU Working Paper No 00/43. Cape Town: Development Policy Research Unit, University of Cape Town. BHORAT, H. 2000b. The impact of trade and structural changes on sectoral employment in South Africa. Development Southern Africa 17(3): September. BHORAT, H., 2003a. The post-apartheid challenge: Labour demand trends in the South African labour market, DPRU Working Paper No 03/82. Cape Town: Development Policy Research Unit, University of Cape Town. BHORAT, H., 2003b. Labour market challenges in the post-apartheid South Africa. Mimeo. Cape Town: Development Policy Research Unit, University of Cape Town. BHORAT, H. & HODGE, J Decomposing shifts in labour demand in South Africa. South African Journal of Economics 67(3): September. BHORAT, H. & LEIBBRANDT, M Modelling vulnerability and low earnings in the South African labour market. Ch. 4 in: Bhorat, H., Leibbrandt, M., Maziya, M., Van der Berg, S. & Woolard, I Fighting poverty: Labour markets and inequality in South Africa. Cape Town: UCT Press: CASALE, D What has the feminisation of the labour market bought women in South Africa? Trends in labour force participation, employment and earnings, DPRU Working Paper No 04/84. Cape Town: Development Policy Research Unit, University of Cape Town. CASE, A. & YOGO, M Does school quality matter? Returns to education and the characteristics of schools in South Africa. NBER Working Paper Cambridge, M.A.: National Bureau of Economic Research. CHAMBERLAIN, D Earnings functions, labour market discrimination and quality of education in South Africa. Masters thesis. Stellenbosch: Department of Economics, University of Stellenbosch. EDWARDS, L Globalisation and the skills bias of occupational employment in South Africa. South African Journal of Economics 69(1): March. DEVEY, R., SKINNER, C. & VALODIA, I Informal economy employment data in South Africa: a critical analysis. Paper to the TIPS/DPRU Forum, 8-10 September, Johannesburg. FALLON, P. & LUCAS, R South African labour markets: Adjustments and inequalities. Discussion Paper 12, Informal discussion papers on aspects of the economy of South Africa. Washington, D.C.: World Bank. FEDDERKE, J., HENDERSON, S., MARIOTTI, M. & VAZE, P Changing factor market conditions in South Africa: The labour market a sectoral description of the period Paper to DPRU/FES Conference on Labour markets and poverty in South Africa held at Johannesburg, November. FILMER, D. & PRITCHETT, L What education production functions really show: A positive theory of education expenditures. Economics of Education Review 18(2): FRYER, D. & VENCATACHELLUM, D Returns to education in South Africa: Evidence from Machibisa Township. DPRU Working Paper No 03/76. Cape Town: Development Policy Research Unit, University of Cape Town. HERTZ, T Upward bias in the estimated returns to education: Evidence from South Africa. American Economic Review 93(4): HOFMEYR, J.F Segmentation in the South African labour market in Paper to DPRU/FES Conference on Labour markets and poverty in South Africa, Johannesburg, November. 13

14 HOSKING, S Rates of return to education in South Africa, Paper to DPRU/FES Conference on Labour markets and poverty in South Africa, Johannesburg, November. KESWELL, M. & POSWELL, L How important is education for getting ahead in South Africa? CSSR Working Paper No. 22. Cape Town: Centre for Social Science Research, University of Cape Town. LUNDALL, P Sector Education Training Authorities and the delivery of training: Preliminary remarks on the new skills dispensation in South Africa. DPRU Working Paper 03/79. Cape Town: Development Policy Research Unit, University of Cape Town. LAM, D Generating extreme inequality: Schooling, earnings, and intergenerational transmission of human capital in South Africa and Brazil. Research Report Ann Arbor: Population Studies Center, University of Michigan. August. MLATSHENI, C. & ROSPABE, S Why is youth unemployment so high and unequally spread in South Africa? DPRU Working Paper No. 02/65. Cape Town: Development Policy Research Unit, University of Cape Town. OOSTHUIZEN, M Expected labour demand in South Africa: Working Paper No 03/81. Cape Town: Development Policy Research Unit, University of Cape Town. POSWELL, L The post-apartheid South African labour market: A status report. Cape Town: Development Policy Research Unit, University of Cape Town. PSACHAROPOULOS, G Returns to investment in education. World Development 22(9): ROSPABE, S Making racial wage relations fair in South Africa: a focus on the role of trade unions. DPRU Working Paper 01/48. Cape Town: Development Policy Research Unit, University of Cape Town. SÀNCHEZ-PÀRAMO, C. & SCHADY, N Off and running? Technology, trade and the rising demand for skilled workers in Latin America. Policy Research Working Paper WPS Washington, D.C.: World Bank. SCHULTZ, T.P. & MWABU, G Wage premiums for education and location of South African workers, by gender and race. Economic Development and Cultural Change 48(2): SOUTH AFRICAN RESERVE BANK Quarterly Bulletin March Pretoria: South African Reserve Bank. SOUTH AFRICA, NATIONAL TREASURY Budget review Pretoria: Government printer. SOUTH AFRICA, STATISTICS SOUTH AFRICA Labour Force Survey Statistical release P0210. Pretoria: Statistics South Africa. TSIKATA, Y Liberalization and trade performance in South Africa. Discussion Paper 13, Informal discussion papers on aspects of the economy of South Africa. Washington, D.C.: World Bank. 14