Document(s) 7 of 11

How the jobs crisis tested the governing alliance

Early in 1998, the Congress of South African Trade Unions (COSATU) fired a shot across the bow of the ANC. It demanded that, in return for continuing to support the ANC election alliance, it, and the other alliance member, the South African Communist Party (SACP) wanted to enter into 'bona fide' negotiations on 'fundamental policies' (Mail and Guardian 16–22 January 1998).

Zwelinzima Vavi, then Deputy Secretary-General of COSATU, announced the confrontation. Vavi is a tall, well-built man, with a forthright demeanour. He singled out the Growth, Employment and Redistribution (GEAR) strategy as the unions' target. 'Any macroeconomic programme that fails to deliver jobs and decent salaries to the poor is wrong,' he said. '[I]t would be immoral for COSATU not to speak out' (Mail and Guardian 16–22 January 1998).

After growing well during the first three years of the new government, the economy slowed down. While exports still grew, the job market stagnated and unemployment rose. In spite of this, the government continued to pursue a fiscal policy that COSATU saw as contractionary. It sought to bring the deficit down by reengineering government, which included cutting back on government employment, holding government wage increases to low real levels, and privatising or restructuring a range of state-owned companies.

The government promised a Jobs Summit during 1998 – where its leaders would meet with those of business and labour and 'the community' to attempt to agree on strategies and programmes that might deliver sufficient new jobs. But the role of the summit was still unclear; visions ranged from a wide-reaching wages and prices pact, to little more than an opportunity for grandstanding by unreconciled social partners.

In spite of the uncertainties, government did not shrink back. President Mandela chose his state of the nation speech early in February 1998 to endorse the 'leaner government' strategy. The opening of Parliament speech is the equivalent of an American presidential state of the union address, or the Queen's speech to the British Parliament. It sets the parameters for the legislative year. Mandela spoke of the need to shed some of the 1.2 million public sector jobs as a key strategy to eliminate government overspending. Minister of Finance, Trevor Manuel, underlined the message. 'There is no room for equivocation' on public sector employment, he said. 'We've got our battle orders to get this thing in line' (Sunday Times: Business Times 8 February 1998).

The South African Municipal Workers' Union (SAMWU) replied angrily. The union's 'anti-privatisation co-ordinator', Maria van Driel, slammed the ANC's endorsement of GEAR at its December 1997 three-yearly national conference as 'immoral'. Van Driel portrayed GEAR as a tool of financial capitalists who promote neo-liberal policies to allow them maximum freedom to make profits. As business taxes are cut to boost profits, fiscal pressures force government to privatise. Privatisation leads to the fragmentation of the working class, and to inferior conditions for workers. So, SAMWU called on COSATU, its federation, to 'maintain its convictions regarding the rejection of GEAR', and not to 'pander to the ANC' (Sowetan 16 February 1998).

Had the economy been growing strongly, SAMWU's appeal might have been seen as special pleading. But the economy was cooling off fast. The onset of the Asian crisis in 1997 had encouraged the South African Reserve Bank (SARB) to maintain excessively high interest rates, and then raise them still further. This attracted a growing pile of short-term foreign credit that was soon to lead banks and ratings agencies to view South Africa as an investment risk. A growing fear that the Asian contagion would not spare South Africa meant that growth rate projections of 2–2.6% early in the year were soon edging downwards. There was no silver cloud. Vavi's contention that the poor could not wait 20 years for delivery began to resonate.

Shortly before Trevor Manuel presented the 1998–99 budget to Parliament, COSATU's Vavi struck again, this time with a critique of the role of Parliament in the formulation and implementation of fiscal policy. He both appealed to the populist sentiments of the ANC parliamentary members to engage with the fiscal process in order to produce a 'People's budget', and chided them for not doing enough. 'Parliament's role is limited to symbolic comments, and the holding of hearings that will not make any difference to the Budget,' he mocked (Cape Times: Business Report 6 March 1998).

During the pre-budget assault, the union and SACP theme was that the GEAR budget deficit target was 'arbitrary' and 'too tight'. Trevor Manuel could not resist reminding his left-wing critics that 'while balanced budgets are the stuff of market orthodoxy now, Karl Marx warned against public debt in Das Kapital' (Business Day 9 March 1998).

But COSATU was not convinced. The next day it announced that it was withdrawing from the parliamentary hearings on the budget. Normally, COSATU, like other interested parties, would submit its views to the parliamentary committee that was responsible for taking the bill to Parliament. It would also expect the former COSATU union leaders amongst the ANC parliamentarians to promote the cause. This time it decided to boycott the committee process to convey its anger about the perceived inability of the committee to have a 'real impact on Budget decisions' (Cape Times 3 March 1998).

In truth, the committee was not entirely powerless. As is the case for all 'money bills' in South Africa, the committee could not amend the budget bill, but could only return it to the drafters in the government. To some degree, then, COSATU was protesting at the party unity and discipline of the ANC parliamentary caucus. Some bills, even money bills, have been blocked or amended by ANC-dominated parliamentary committees. But for the ANC-dominated parliamentary committee to challenge the budget would have been tantamount to a vote of no confidence in the Cabinet and the ANC leadership.

Union anger continued to smoulder in the run-up to the COSATU and SACP congresses in mid-year. Sometimes this was reflected in documents that were circulated in preparation for the Jobs Summit. For example, the first labour input to the Presidential Jobs Summit called on government to sustain current levels of public sector employment by 'lower(ing) interest rates, restructuring the taxation system and reducing the burden of apartheid debt'. It also called on government to reverse trade policy reform, and to pass further laws to regulate the labour market (COSATU et al. 1998a).

Tripartite meetings between the ANC, COSATU and the SACP failed to reduce the temperature of the debate. In draft resolutions prepared for a mid-year central committee meeting, COSATU reaffirmed 'its rejection of GEAR and the government's macroeconomic practices'. COSATU revived its own pet obsession saying: 'We reject the obsession around deficit reduction' (Financial Mail 19 June 1998). (The government had proposed a gradual reduction of the fiscal deficit from 5.1% in 1996 to 3% in 2000.) The SACP documents for its July congress adopted a similar tone. Commenting on these developments, political reporter, Justice Malala, wrote in the Financial Mail: 'The divisions over GEAR are the deepest that the tripartite alliance has ever experienced' (Financial Mail 19 June 1998).

In his speech to the COSATU meeting, Deputy President Thabo Mbeki, who had been elected President of the ANC in December 1997, acknowledged the perilous state of the alliance: '[W]hen we speak of this strategic alliance, are we speaking of something that continues to exist, or are we dreaming dreams of the past?' he asked. 'Does a Congress movement still exist[?] Do we have the right to call one another Comrades signifying a commitment to our being fellow combatants for liberation... or are we calling one another Comrade simply because we are no longer used to calling one another Mr, Mrs or Miss' (Mbeki 1998b: 2).

President Mandela was equally challenging at the SACP meeting a week later. Mandela departed from his prepared text when he heard SACP leader, Charles Nqakula, and a COSATU representative criticise the GEAR programme. Mandela thundered: 'GEAR, as I have said before, is the fundamental policy of the ANC. We will not change it because of your pressure. If you feel you cannot get your way (then) go out and shout like opposition parties. Prepare to face the full implications of that line' (Business Day 2 July 1998).

Then Mandela called on the SACP to 'choose a role you want to play'. After a polite comment from the chairperson of the session, the hall sat in stunned silence as Mandela slowly departed (Business Day 2 July 1998).

One reason for Mandela's anger was the fact that the Asian contagion seemed to entail potential danger for South Africa. Standard and Poor's risk study in July 1998 placed South Africa second riskiest, after Indonesia, on a list of 10 emerging economies (Business Day 17 July 1998). Though this was soon shown to be rather inaccurate by the Russian and Brazilian crises, it demonstrated the vulnerability of the South African economy. In this context, the posturing of COSATU and the SACP seemed like disloyalty to the ANC.

The looming Jobs Summit took on an added significance. Not only did it need to come up with feasible strategies to address the unemployment crisis, it also had to show the world that South Africans were not fragmenting, but were working together for the future of the whole country. This was how the ANC leaders and the government increasingly saw the Jobs Summit.

To achieve this objective would require two necessary conditions: firstly that the parties agree on some more or less significant policies or programmes that could make a dent on unemployment, and secondly, that the ANC and its alliance partners patch their wounds through some symbolic re-engagement. It also required a degree of agreement amongst the social partners about the causes of the unemployment problem, an issue that had long been the subject of an acrimonious debate.

In the end there was a Jobs Summit in October 1998, with a set of agreements. But the limited nature of the agreements and the poor follow-through after the summit showed how far apart the major economic actors were, both in their analysis of the causes of unemployment and on the appropriate solutions. Trade-offs to reach agreement left the country with a lowest common denominator outcome. For example, business agreed not to raise questions about labour laws and regulations, and labour agreed to lay off its criticism of the GEAR strategy. One of the few substantive outcomes was an agreement to develop a programme to support the purchase of locally made products and services that ultimately became the Proudly South African campaign, but even this took years to implement because of residual disagreements about the orientation of the programme.

What did the Jobs Summit achieve? One important achievement was that it showed that when it came down to it, business, labour and government could at least achieve a show of unity to the rest of the world. Another achievement was that it encouraged business to come up with a novel and exciting programme to establish a business trust to support economic development programmes, but the trust was not a formal outcome of the Summit.

A similarly indirect outcome of the Jobs Summit was that its mediocrity suggested to then Deputy President, Thabo Mbeki, that new channels of communication should be opened up between the major economic stakeholders. One of the weaknesses of Nedlac was that it was becoming a meeting point for middle-level leaders from business, labour and government – especially the business contingent. These were often representatives with vague mandates and no authority to improvise. The senior leadership of business, government and labour seldom met, and when they did they focused on procedure rather than substance. Mbeki established channels of communication directly with leaders of business and labour – eventually four economic 'working groups' were established under Mbeki after he became president in 1999. They are: a Big Business Working Group, a Black Business Working Group, a Trade Union Working Group and a Commercial Agriculture Working Group (for black and white farmers). Each of these groups of leaders meets with the President and an appropriate group of Cabinet Ministers and Deputy Ministers, plus some senior staff, usually twice annually. Twice a year there is usually a Joint Working Group meeting where all four groups meet together with the President, ministers and a few top officials.

The meetings are essentially opportunities to exchange views and voice concerns – they are not sites for negotiation, and they are private. This has allowed for unusually free and frank exchanges. The Towards a Ten Year Review publication of the government remarked that one of the constraints on investment and growth in South Africa was the lack of trust between the economic and political elites because of the historical reality that most business leaders were white and most political leaders were black, and that they shared very few experiences (Policy Co-ordination and Advisory Services, The Presidency 2003). This was one of the key reasons for the establishment of the working group system. The other, related reason for the working groups is that the President clearly believes that there are structural impediments to growth and development that can only be confronted in frank exchanges between the major economic stakeholders. One of the slogans of Mbeki's ANC in the 2004 election campaign was 'A people's contract for growth and development'. Mbeki and his party clearly believe that in order for the ceiling on growth to be breached, more and deeper mutual commitments are required between the social partners. This is why Mbeki called for a Growth and Development Summit, which took place in June 2003. The key targets of the Growth and Development Summit, like those of the Jobs Summit in 1998, were to roll back unemployment and poverty.

How bad is the unemployment problem?

South Africans believe that the biggest economic problem in South Africa is unemployment. Opinion surveys show an overwhelming and still rising consensus that not 'poverty', 'low wages', 'housing' or 'inequality', but specifically 'unemployment' is South Africa's key challenge. The unemployment rate is very high in South Africa.

Unemployment is generally seen as a synonym for poverty. Certainly, poverty surveys in South Africa have shown that the poorer you are, the more likely you are to be unemployed, and that poorer households have a higher percentage of members who are unemployed (May 1998: 81–86). So, unemployment and poverty have become almost indistinguishable in South Africa.

Just how large the unemployment problem is, is not a simple question. An International Labour Organisation (ILO)-sponsored study proposed 14 reasons why employment estimates produced by the South African government might be wrong – mostly under-estimating employment levels (Standing et al. 1996: 68–71). Unemployment statistics were a terrain of political debate during the apartheid years, and remain politically charged.

In response to this challenge, the South African government's revamped statistical agency – Statistics South Africa – overhauled the employment measurement machinery in the late 1990s. Household surveys were introduced to compliment the census data and the firms-based surveys. The October Household Survey (OHS) was introduced in 1995 and became the Labour Force Survey (LFS) in 2000. The estimate of unemployment for September 2004 was 26%, which is considerably worse than the level of about 17% recorded in 1995. This is according to the 'official definition', which uses the convention that a person is unemployed if he has not worked in the previous seven days, was available to take up work within a week of the interview, and had sought work or self-employment in the past four weeks. If one uses an 'expanded unemployment rate' that does not require that the person sought work in the past four weeks, the rate rises to 40%, compared with 27% in 1995 (Statistics South Africa 1998; LFS, various).

Unemployment is very unevenly distributed. A very high proportion of young people are unemployed – 75% of the unemployed are under the age of 35. This is partly a result of queuing, poor, young people slowly making themselves employable or finding self-employment, and partly a result of the likelihood that more people over 35 no longer seek work, though they are unemployed. Unemployment weighs more heavily on 'African' South Africans, compared with 'coloured', 'Indian' and 'white' South Africans who have respectively increasingly lower levels of unemployment. Historically, unemployment has been higher in the rural areas of South Africa. This was an anomaly by all standards, the result of the influx control policies of the apartheid government. Part of the purpose of apartheid was the removal of black unemployed people from politically vulnerable urban centres to remote rural areas. Many, especially the poorly educated, still find it very hard to escape what are effectively rural poverty ghettos. The remoteness and backwardness of the former 'African reserves' raises the costs of job seeking to a level that many poor households can barely afford. This situation rapidly adjusted in the post-1994 period, though there is still an unusually high level of unemployment in South Africa's rural areas (Statistics South Africa 1998, 2003).

Does this mean that South Africa experienced jobless growth? Growth has averaged about 3% in the post-1994 era, but what happened to employment? This is a matter of some controversy, as the employment data doesn't offer definitive answers. The definitive employment series, the bi-annual LFS, began in 2000. Though it picked up where the OHS (1995–99) left off, there are discontinuities in some definitions. For example, the definition of employment in the informal sector changed a little with the LFS, and some analysts argue that the early OHSs under-counted employment in some sectors. There are also weighting issues, but these are relatively easy to accommodate.

The alternative measures are not any better. The Survey of Earning and Employment, a firm-based sample survey, deteriorated during the 1990s, and when the sample was re-based in 2002 to accommodate structural changes in the economy, the index of employment increased by 40%. The population census has very limited questions on employment, and, in this respect, is seen by the statistical service as no more than a way of roughly verifying the employment data of the LFS.

The lack of a consistent employment series going back beyond 2000 made the issue of employment statistics politically controversial, both between the ANC government and the opposition, and between the ANC and its ally, COSATU.

According to various OHS/LFS surveys, about 9.5 million people were employed in 1995, with about 8.3 million in the formal sector. By 2003, about 11.5 million people were employed in South Africa, with about 8.2 million in the formal sector. The increase in employment came largely from the domestic workers (0.3 million), subsistence farmers (0.4 million), and the rest of the informal sector (1.4 million). However, as subsistence farmers were not counted in 1995, this number should be discounted as a trend. Within the formal sector there were significant shifts, with commercial agriculture falling from 1.1 million in 1995 to 0.9 million in 2003, the public sector falling from 1.7 million to 1.6 million, and the private non-agricultural sector rising from 5.5 million in 1995 to 5.8 million in 2003. About 20% of employees were temporary, casual or had fixed-term contracts.

In short, according to the OHS/LFS surveys, the number of employed people rose by something in the vicinity of 2 million people, but it could be as high as 2.2 million, or as low as 1.5 million. Most of the new jobs in the 1990s were in the informal sector, and, especially after 1999, a rise in formal non-agricultural employment compensated for a fall in farm workers and public servants.

Though it is not desirable that so many of the new jobs came from the informal sector, at least growth in private sector non-agricultural jobs compensated for the fall in farm employment and public servants – the latter a result of the belt-tightening exercises of the late 1990s. Another important trend is the revival of the private sector job market after the pain of restructuring in the 1990s. While private formal sector non-agricultural jobs fell from 5.5 million in 1995 to 5.0 million in 1999–2000, they then grew to 5.8 million by 2003. By contrast, almost all of the growth of informal sector jobs took place before 2000, after which this sector was virtually stagnant.

The mean rate of overall job growth was about 2.1% per year – with a mean economic growth rate of about 2.8% per year. This suggests a coefficient of 0.75 between the rate of growth and the rate of job creation. This is a relatively elastic employment response. Assuming that elasticity is in fact 0.75, a steady growth rate of 4% would yield a job creation rate of 3% per annum, or about 345 000 net new jobs per annum (Bhorat 2003). There are substantial differences in the employment elasticity of different sectors, so the sectoral composition of growth would be significant. Also, a significant proportion of these new jobs may be in the informal sector, and some may be temporary, casual or fixed-term contracts.

Even though employment grew between 1995 and 2003, the number of unemployed and the rate of unemployment grew steadily. How is this possible? How can it be that unemployment grew in spite of the fact that job creation grew faster than the population growth rate?

The answer is that the economically active population (EAP) grew at an exceptionally fast rate over the same period. The EAP consists of those people who are either employed or who are seeking employment, and are aged between 15 and 65. The EAP grew from 11.4 million in 1995 to 16.8 million in 2003. This represents a growth rate of about 5% per year, which is well above the population growth rate (about 1.6% in 2003). Why should the EAP have grown so exceptionally fast over this period? Studies suggest that the main factor is that many African women, who defined themselves outside of the job market in 1995, now define themselves as employed or seeking work. This is a result of migration from the rural areas, a degree of social liberation of women, and the fact that a very significant proportion of the new jobs created since 1995 were open to and attracted African women. Another factor could be the rising education levels (Posel and Casale 2003).

So we have a paradoxical situation. Because of the very rapid growth of the EAP (people defining themselves as employed or seeking jobs), the fact that we have had job growth has not prevented the unemployment rate from rising rapidly to over 30% in 2003 before falling back to 26% in 2004.

What this suggests is that the social and demographic changes we have experienced since the mid-1990s have made the task of reducing the rate of unemployment even more difficult than we expected. A key question for which there are no definitive answers yet is: will the labour force participation rate continue to rise, and if so how quickly? The trend-line for labour force participation seems to have flattened in the 2000s at around 68%, but it is still relatively low compared with developed countries.

What causes unemployment in South Africa?

We should explore the 'market clearing' approach first. It dominates international perceptions of South Africa for those influenced by the Bretton Woods institutions – in a sense the economic policy elite.

The inflexible labour market

It has been argued for a while now that the labour market in South Africa is less flexible than it should be, as a result of political and institutional factors. Some of these factors, such as racial and gender prejudice, are relics of the past. Some, like revised labour laws and the role of trade unions in economic management through Nedlac, are products of the political transformation. The leading role of trade unions in the struggle for democracy meant that they won victories that sheltered the determination of wages for black workers from the unforgiving heat of the markets.

South African trade unions have a strong egalitarian streak, and tend to strive for the narrowing of wide wage gaps, so that unskilled workers got the most benefit. The new government followed suit by raising the wages of the lowest ranks of government employees most rapidly. So, the loss of 'flexibility' or responsiveness to markets became most significant for the unskilled sector.

Peter Fallon of the World Bank and Robert Lucas from Boston University reviewed the evidence of market clearing problems in a 1998 study. They did not blame unemployment simply on the inflexibility of the labour market, but they saw it as a key factor.

Fallon and Lucas showed that the wages of white workers were more responsive to changes in labour market conditions than the wages of black workers. In other words, if the wage employment of Africans and whites each fell by 1%, real African wages would fall by only 0.22–0.27%, but white wages would fall by between 0.93–3.18%. This shows, they argued, that white workers' wages react more efficiently to changes in labour market conditions. The data also showed that unskilled wages adjusted most slowly to changes in market conditions. The rise in unemployment levels from 1980–94 should have led to wages falling for African workers, but they rose by 20% (Fallon and Lucas 1998: 12–14).

The main reason for African wages reacting perversely in relation to levels of unemployment is the strength of the black trade unions, argued Fallon and Lucas. This argument is supported by the finding, according to one survey, that a unionised unskilled worker would be paid a remarkable 71% more than an non-unionised worker with the same lack of skills (SALDRU 1993).

It is not really a matter for dispute that unskilled and semiskilled workers in the formal sector in South Africa are paid relatively more than their counterparts in some other parts of the world, including the surrounding countries of southern Africa. There is evidence of this in all the major studies drawn on for this chapter, whether written by or for the left or the right (Fallon and Lucas 1998: 12–14).

In my own work in the 1980s I found that workers employed in Bantustan factories outside of the South African wage bargaining system, were paid around half as much as workers in similar employment in a nearby town lying outside the Bantustan, and around one quarter of the average wage for Africans in the relevant sector in South Africa (Hirsch 1987: 268–271).

In an informal survey in 1994, I found that basic operatives in the textile and metal working industries were paid about 15 times more in South Africa than in Tanzania. But the wage level did not mean that South African workers had a standard of living 15 times better than their counterparts in Dar es Salaam – apartheid led to excessively high living costs for workers in South Africa, especially transport costs as a result of population location policies, as we shall discuss shortly.

The big question, though, is whether, and to what extent, the high level of unemployment amongst unskilled workers can be attributed to labour market inflexibilities and distortions. One example of the explanation of unemployment that emphasises the inflexibility of labour markets for the unskilled is found in the 1996 South Africa Foundation Growth for All document, and in several iterations of the International Monetary Fund's (IMF) annual Article IV South Africa staff report (for example, IMF 1998: 19–21).

The South Africa Foundation's Growth for All document gave the impression, deliberately or accidentally, that the main cause of unemployment was the distorted labour market resulting from union pressure and government regulation. For this reason, the report emphasised the need for restructuring the labour market as a 'two-tier market' and seemed to underplay other possible corrective actions (South Africa Foundation 1996: chapter 6).

This was an analytical and a political error. Political, because it set up a conflict between other social forces and the unions which, with their backs rubbing up against the proverbial wall, linked arms and marched forward. It forced the ANC, if temporarily, into the arms of the unions. It led to damage in the social fabric that was only partly patched up nearly three years later at the Jobs Summit.

But why was it an analytical mistake? The main problem is that it ignored the history of South African unemployment. According to Charles Simkins, an authoritative analyst of the history of unemployment in South Africa, unemployment probably began to escalate in the 1960s, during an economic boom, and in a period when black trade unions were banned and black workers were severely oppressed. Simkins found that black (African) unemployment, broadly defined, already stood at about 12% in 1970 and rose to slightly more than 2 million, or 21%, by 1981 (Simpkins 1982). According to the official Current Population Survey, narrowly defined unemployment was already as high as 13% in 1977, and another official government report put broadly defined unemployment at 30% in 1980 (Archer et al. 1990: 166–167).

Structural unemployment

Black trade unions only began to re-emerge in the mid-1970s, after being crushed by the apartheid government in the early 1960s. Trade unionism for Africans was only re-legalised in 1979, having been excluded from the bargaining system in the 1950s. Therefore black trade unions cannot be blamed for the emergence of high levels of unemployment in modern South Africa. It is even hard to blame high wages, as black wages in the 1960s, when unemployment emerged as a significant phenomenon, were at historically low levels.

This realisation led to South African analysts resorting to a concept of 'structural unemployment' that differed from the in-flexibility argument. 'Structural unemployment' explained David Lewis, 'refers to a deep-seated disjuncture between the growth and structure of the national workforce on the one hand and the growth and structure of economic activity on the other hand' (Lewis 1991: 246).

One explanation, used by Alec Erwin before he left academia for the trade unions, and many years before he became a Cabinet Minister, was that capitalism in South Africa was incomplete – it was a stunted, colonial version of capitalism that could not generate balanced growth or adequate employment (Erwin 1978). The implication was that South African production should move into fields controlled by the advanced capitalist countries – that South Africans should not be confined to mining and exporting gold and making consumer products for the domestic market only.

A later version was that South Africa was unable to make a transition to a more modern form of capitalism than stunted 'racial fordism' because of the nature of white racist political rule (Gelb 1991a). Black progress in the job market, into skilled, professional and managerial roles, was incompatible with white minority rule. This had elements in common with Erwin's approach, in that both suggested that the anti-colonial struggle had to be completed for South Africa to make the transition to a healthy growing economy.

White rule and its opposition to the development of a competitive black middle class in the modern urban sector set up obstacles to black business (see Chapters One and Six), demolished black education, and decimated industrial training systems as white artisans moved into managerial positions. A reasonably good apprenticeship system for whites was virtually destroyed, and not replaced. The result was that an annual global survey of competitiveness placed South Africa last in 'investment in people' virtually every year since 1994 (Cape Times: Business Report 18 September 1998; Garelli 1998).

Another approach was to show how the incentive structure of the South African economy favoured capital-intensive investments in the 1970s and 1980s. The structure of protection, subsidies on investment, and low or negative interest rates conspired against the employment of unskilled workers. As Merle Lipton put it:

This distortion of relative factor prices (making capital artificially cheap and labour dear) led to a costly and inefficient development path. It encouraged capital intensity, thus increasing the demand for scarce resources such as capital and skill, while discouraging the use of what by the 1960s had become a surplus of unskilled labour, so contributing to growing unemployment (1986: 244).1

Certainly, South Africa experienced very low or negative interest rates (tilting investment towards capital-intensive projects) for most of the latter half of the 1970s and most of the 1980s. One could extend this argument into a critique of certain investment incentives that encouraged capital-intensive projects, such as Section 37e of the Income Tax Act in the early 1990s, or of the investment priorities and organisational bias of the Industrial Development Corporation (IDC).

Trade unions and a government that was soft on workers did not create the unemployment problem. Apartheid's social and economic policies were the main cause. However, there is a risk, identified by not only business, the World Bank and the GEAR strategy, but even by the government's own study of poverty and inequality in South Africa, that emerging rigidities in South African labour markets could block investment and employment opportunities.

The GEAR strategy pointed to the need for 'a more flexible labour market'. It indicated that wage settlements and agreements should 'be sensitive to regional labour market conditions, the diversity of skills levels in firms of varying size, location or capital intensity, and the need to foster training opportunities for new entrants to the labour market'. Industrial agreements 'which reach across diverse firms, sectors, or regions should be sufficiently flexible to avoid job losses and should be extended to non-parties only when this can be assured'. The GEAR strategy also argued against the establishment of a national minimum wage, but rather that minimum wages should be set appropriately for sectors and areas and should take into account 'the potential for employment creation and the alleviation of poverty' (Department of Finance 1996: 18–19).

The GEAR report pointed to three key problem areas: the periurban and rural regions where unemployment is high; small firms, which cannot yet support top-class conditions; and the never-been-employed youth. As the Poverty Report notes, a more flexible labour market will help draw in the unemployed, though it must be tempered with as much job security as possible, and through government improving the social wage – improved access to public services and facilities for the poor (May 1998: 90).

Perhaps the most important regulatory issues in this respect concern: the capacity to extend bargaining council agreements to non-parties (for instance, to small towns); the powers of the Employment Conditions Commission (for workers not represented in bargaining councils); and the wage-setting example of the government sector. All three should be carefully managed so as to avoid bidding unemployed people out of the system.

Trade liberalisation and unemployment

There was heated debate over the effects of the trade liberalisation programme on employment. In an early submission for the Jobs Summit, the trade unions urged:

Stop the damaging effects which trade liberalisation has on employment. Where South Africa has lowered individual tariffs to below GATT commitments and these have resulted in job losses, these must be increased up to the GATT binding rate (COSATU et al. 1998a: 2).

But the evidence on the relationship between trade liberalisation and employment is hard to evaluate, especially on a sector-by-sector basis. Johannesburg-based economist Simon Roberts noted that there is no workable 'general explanation' on the effects of the tariff reform programme on growth and employment, and that 'recent developments can only be understood through an examination of sub-sector specific factors together with industrial structure'. His sectoral study failed to establish consistent results. In some sectors exports and employment increased, in others both decreased, while in some, exports grew but output and employment stagnated or declined. For manufacturing overall, exports grew steadily, but employment stagnated or declined (Roberts 1998: 26).

A trend that has been noted is that much of the strength in manufactured exports has come from the more capital-intensive sectors of manufacturing, while a significant proportion of the increase in imports is in labour-intensive sectors (Bell and Caettaneo 1997). However, another broad trend, detected by an ILO study, is that jobs have generally been gained in the non-tradable and import-competing sectors, and lost in the export sectors in spite of output and export growth in these sectors. It also found that sectors with significant tariff cuts performed best with regards production, exports, employment, wages and productivity. Part of the explanation is that export sectors have been forced to streamline in the light of a higher competitive temperature and the disappearance of export subsidies. More recent academic work has come to similar conclusions (Hayter 1999: 2–6; Edwards 2001: 40–72).

It is possible to come to two alternative policy proposals. One would be to suggest that liberalisation is a mistake if your economy is not sufficiently competitive in labour-absorbing industries – the problem is that only capital-intensive sectors are reaping the export benefits of greater efficiency (Bell and Caettaneo 1997). The alternative would be to address the problems in the labour market, to improve the supply of cheaper, sufficiently skilled labour, and to remove obstacles to investment in the labour-intensive sectors so that efficiency gains could be exploited in employment-absorbing sectors (Nattrass 1998; Tsikata 1998).

Bear in mind that some of these studies misunderstand the main purpose of trade reform, which is to remove price distortions in the domestic market. As Minister Alec Erwin argued in his presentation to the Jobs Summit, the tariff reform programme is 'designed to induce previously protected enterprises to become more efficient and to reduce domestic cost levels so that they are closer to general world cost levels' (Erwin 1998: 5). This makes the economy as a whole more competitive, and will encourage investment and job creation, but will not necessarily produce predictable results on a sub-sector by sub-sector basis, unless the analyst is able to account for all the particularities that frustrated Simon Roberts in his study. There will also be lags between negative and positive effects in some sectors. A U-shaped labour demand curve should be the anticipated outcome of a significant trade policy reform.

The floor price of labour in South Africa

But there is an underlying and rather obvious point that must be made. There are very few labour-intensive manufactured tradables that South Africa is likely to be competitive in. The competitive labour-intensive sectors will be those where South African producers are able to exploit an underlying advantage, such as cheap or special access to natural resources, or a preferential market arrangement. To put it another way, no matter how much the labour market is made more flexible or the currency is depreciated, South Africa will never have tens of thousands of workers making Nike shoes for export, like Vietnam or Thailand.

Why is this? One reason is the strength of the trade unions combined with the country's human rights culture. Tables 5.1 and 5.2 indicate the growth of worker rights, and the continuation of worker militancy in democratic South Africa. Another reason for relatively high wages is the high cost of living in South Africa. Apartheid raised living costs for all South Africans, especially the poor, one of the main costs being apartheid-determined residential locations (Department of Transport 1998). The working poor were located miles from their potential places of work, and often equally far from commercial and public services. Public transport systems had practically collapsed by 1994, and some of the new distant locations were not served at all by public transport. Though public transport services have begun to improve under the new democratic government, pressures to reduce the government deficit forced it to reduce subsidies that might have helped the poor and working poor use public transport. The deterioration of access to public services such as education, heath, and social security for Africans under apartheid meant the diminution of the social wage.

Moreover, in South Africa the earned wage represents a much higher proportion of most households' income than is the case in some surrounding African countries, where many workers have non-wage incomes often deriving from their rural land-holdings. In South Africa the expulsion of Africans from the land was far more complete under a vicious form of settler colonialism than it was in most African countries where settler interests were weaker. South Africa has neither significant agricultural subsistence income nor a sufficient social wage for worker households. If living standards were actually 15 times higher in South Africa than in Mozambique and Tanzania, illegal immigration would be much higher than current high levels.

It will take time to bring down the cost of living and raise the social wage for the poor and working poor in South Africa. Trade liberalisation has already helped to reduce the cost of living of the urban poor, and there is further downwards potential for the price

of wage goods in southern Africa. The social wage is improving as the poor and working poor get better access to health and education services. In the future new urban residential and transport policies could help too. As the cost of living and the social wage of the working class improves, and as the number of suitably skilled workers increases, competition for employment should prevent the cost of employment from rising too fast. In the meantime, though, it is important to ensure that opportunities for employing the poor from what President Mbeki calls the second economy (the marginalised poor) are not forgone due to inappropriate regulatory constraints.

The skills challenge

The skills crisis in South Africa in some ways resembles the challenges facing many other countries, but in several important ways it is unique. What it has in common with the rest of the world is that the structure of the economy has changed more quickly than the institutions that impart education and skills. The decline of employment on the mines and farms, and in some other sectors, and the rapid rise of employment opportunities in the service sector and more skilled jobs in the manufacturing sector, was as dramatic in South Africa as in any other country. Employment in mining fell by about 40% in the 1990s, while employment in the financial services sector rose by 300%, to illustrate the shift (SARB 2005: table S-138).

More specific to South Africa are the historical reasons for the weakness of the supply of skilled workers. The damage to the general education system done by apartheid is briefly described in an earlier chapter. The system of industrial training was equally damaged by apartheid. The job colour bar meant that there was no point in training Africans as artisans or professionals. It was illegal to award apprenticeships to Africans. In addition, the system of industrial training was falling behind modern industrial trends. This was recognised by training experts in the old regime, and was detailed in reports commissioned by the National Manpower Commission and the National Training Board published from the early 1980s to the early 1990s. The problem was that the continuation of apartheid meant that the reforms proposed by the Commission were difficult to implement. The education system was fragmented, the training system was decentralised and un-integrated, and the colour bar still ruled (Kraak 2004a).

The result was that the number of people trained in enterprise training programmes fell dramatically, by more than 75%. Whereas in 1986, 736 581 people were exposed to industrial training, of whom 29 826 were apprentices in training, by 1998 the total number of industrial trainees had fallen to 152 870, of whom only 16 577 were trainee apprentices. The system of industrial training was disintegrating, and was not yet replaced by a more modern and non-racial system (Badroodien 2003).

The result is what is now widely termed a mismatch between the supply of labour (unskilled) and the demand for labour (skilled) in South Africa. Iraj Abedian, then chief economist at the Standard Bank, estimated that 300 000 vacancies remain unfilled in South Africa because of the shortage of suitably skilled labour (Abedian 2004). It can also be speculated that South Africa would have received considerably more domestic and foreign investment had there been sufficient numbers of suitably skilled people.

The government's response had its origins in COSATU policies that sought industrial training to support multi-skilling, and the broader development of workers. The driver of the Skills Development strategy in the Department of Labour after 1994 was Adrienne Bird who, before 1994, led the discussion on industrial training in COSATU as an official in the Metal and Allied Workers' Union (MAWU). A human resources green paper surfaced in 1996, followed by a Skills Development strategy green paper in 1997 and a Skills Development Act in 1998. Finally, in 2001, the Minister of Labour launched the national Skills Development strategy (Kraak 2004b).

One reason why the process was so extended (Adrienne Bird published a paper outlining the proposals in 1991) was because of the deeply consultative mode of policy development adopted by the labour department. All stakeholders had to commit themselves to the policy – unions, employers, government and training providers. Another factor was the reluctance of the National Treasury to agree to a levy-grant funding system. But, eventually the plan was ready to roll out.

The funding system eventually agreed to by all was a payroll levy of 1%, of which a small amount goes to administration, about a fifth goes to a 'national skills fund' for training for SMEs (small and medium enterprises) and the unemployed, and the remainder can be returned to the employer as a grant on completion of accredited training programmes. Hence the use of the term: 'grant/levy'. Clearly the grant is intended as an incentive to encourage employers to mount suitable programmes. The National Treasury added a small additional tax incentive for the training of unemployed potential workers.

Accredited programmes include 'learnerships', which are an evolved form of apprenticeship aimed at new employees or the unemployed, and skills programmes, which are shorter and focus on specific skills. All programmes are required to combine theory with practical experience in the workplace. All programmes must be accredited by the South African Qualifications Authority in terms of the National Qualifications Framework. The system is administered by 23 Sector Education and Training Authorities (SETAs), while the training programmes themselves can be undertaken by employers or other accredited training providers.

The annual income of the system from the levies paid was about R4 billion in the 2004/05 financial year. It is an ambitious and complex programme. It required the establishment of a large number of new institutions with huge responsibilities. The national Skills Development strategy is undoubtedly the most ambitious and complex of all of the new South Africa's social development programmes. It is not surprising that the system took a long time to implement. Initially the programme has been more successful with larger firms and in higher skills bands, and has not been very successful with smaller firms and illiterate, unskilled or unemployed workers. The tax incentive has not had a significant effect, possibly because it was set at too low a level. Some of the new institutions have functioned well, but some of the SETAs have failed badly. The Labour Department was forced to amend the skills law in 2003 to give the Minister of Labour more power to intervene in underperforming or failing SETAs.

But the Labour Department and the Minister have worked hard to get support for all aspects of the programme. In the context of high levels of unemployment, the issue of learnerships for the unemployed was emphasised, with the ambitious target of 80 000 people in learnerships by March 2005. Government sought and won the commitment of business and labour to support the achievement of this target, and an intermediate target was written into the Growth and Development Summit agreement.

The way forward: A real social contract?

There are a host of underlying reasons for high levels of unemployment in South Africa. One was the colonial form of capitalism which, for a lack of democracy, was not impelled towards full employment. Strong national industrial unions, which steeled themselves in the struggle against apartheid, seized and defended their turf vigorously. The lack of a bond of trust between the political and economic elite is a further key element of the political economy of uneven or incomplete capitalist development in South Africa. On top of this, poor economic policies, isolationist strategic decisions, and the effects of globalisation on the worldwide distribution of investment resulted in a kind of dual economy in South Africa, not based on the long-term expansion of the domestic economy, the spread of employment, or the rise of incomes.

President Thabo Mbeki has used the metaphor 'two economies' to describe this situation (ANC Today, weekly e-mail newsletter 24 August 2003). One part is the modern, globalised, industrial sector; the other is the place occupied by the households without a breadwinner in full-time employment, or without productive assets who live on the margins of the first economy. When measured for the year 1999, 3.7 of 11.4 million households fell below the poverty line, about one third of households.2 The degree of inequality, before social transfers are included, remains very high, with a Gini co-efficient of about 0.59 in 2000 (Policy Co-ordination and Advisory Services, The Presidency 2003: 91). The extent and deep entrenchment of this inequality is an obstacle to development.

Overcoming this legacy required more than the half-hearted 1998 Jobs Summit. It required a concerted and united social effort. This was why in his state of the nation address in February 2002, President Mbeki called for a Growth and Development Summit (GDS). The idea was to move beyond the lowest common denominator outcome of the 1998 Jobs Summit. Perhaps he believed that the relationship-building processes of the intervening years would support a more substantive outcome.

The date for the summit was eventually set for 7 June 2003. As in the case of the 1998 Jobs Summit, Nedlac convened the GDS, but there was a subtle difference – as the President had initiated it, this became a 'presidential' summit or 'the President's summit'. This had the effect of keeping the initiative with government, rather than with Nedlac, partly because of the disappointment of 1998, and because Nedlac's performance in some recent industry sector summits suggested that it was not as effective an organisation as it had been in the 1990s.

In the negotiations process, but particularly at the Summit itself, it was evident that there had been some positive shift of tone in the relationship between government, business and labour since 1998. The culmination of several months' intensive negotiations was an agreement that differed from the Jobs Summit agreement mainly in terms of its breadth. The lengthy agreement focused on four major themes: 'More jobs, better jobs, decent work for all; addressing the investment challenge; advancing equity, developing skills, creating economic opportunities for all and extending services; and local action and implementation' (Department of Labour 2003b: 26). In each of these areas broad and detailed commitments were made implicating one or several parties. The two largest and most concrete agreements were support for an expanded public works programme driven by government, and for the learnerships being rolled out by the Department of Labour and the SETAs. On the learnerships, specifically, government and business agreed specifically to register 72 000 learnerships by May 2004.

Supervision over the implementation of the agreement became the responsibility of Nedlac's executive council. Initially the pace of implementation of most of the agreements was disappointing. It is possible that the election scheduled for April 2004 could be blamed for distracting the parties, but this may not be a sufficient explanation. The only programmes where clear progress was evident were in the learnerships effort, and in the government's Expanded Public Works Programme (EPWP).

The EPWP was developed by government in response to the jobs crisis in the second economy. The philosophy behind it was that there were unemployed people in the second economy, usually younger and more educated, who had the chance to migrate from the second economy to the first economy, and that they lacked experience, information and skills. The EPWP has four elements: labour intensive construction projects; conservation programmes; social service programmes; and economic sector interventions. One reason for getting support for the EPWP from the GDS was that government wanted an arrangement on employment conditions that did not comply with the outcome of regular national agreements, first agreed to at the 1998 Jobs Summit, to be extended to beneficiaries drawn in under the EPWP.

The first element, labour-intensive construction, essentially earmarks a small but increasing portion of the government's growing infrastructure programme for implementation through labour-intensive techniques. The Department of Public Works identified categories of projects that are suitable for manual labour rather than using 'yellow machines', and developed suitable implementation protocols. Projects are suitable for manual labour implementation when the manual labour system produces quality outcomes similar to mechanised systems, and at no extra cost.

In order to access a portion of their Municipal Infrastructure Grant (MIG) or Provincial Infrastructure Grant (PIG), the relevant agency has to assure National Treasury and the Public Works Department that it will adopt these suitable manual labour techniques. The initial portion of the MIG and PIG earmarked for labour-intensive construction is quite small, but it will grow over the five-year duration of the current programme. It is expected that by the end of the five-year period, over 700 000 people will have been employed in such labour-intensive projects, for a period averaging four months. During this time, the beneficiary, who will be drawn from the locality of the project, will also receive formal training, not necessarily specific to construction, and will be supplied with labour market information. As this programme is inspired by an existing successful provincial project, expectations of success are high.

The conservation programmes are similar to the labour-intensive construction programmes in many ways, but they are usually of somewhat longer duration. They are also based on existing successful projects.

Perhaps the most innovative element of the EPWP is the extension of the idea of public works to social services. It was recognised that people in the second economy are greatly in need of certain social services such as early childhood development (educare for pre-schoolers) and home-based health care, especially with the burdens of AIDS, TB and malaria. Government could not afford to extend these services as part of the formal public service, so the idea emerged that such services could be extended through semi-skilled health and educare workers, employed by sub-contractors to government. Again, this is not a new modality in South Africa, but one that could be significantly expanded. In this case, the duration of contracts would be up to two years, and the training support would aim at producing a strong group of proto-nurses and proto-teachers who might later be absorbed into formal teaching and nursing.

The economic component of the EPWP is essentially various forms of support for micro-enterprise development including credit and training.

In all, the EPWP was planned to include one million direct beneficiaries in the five-year period between 2004 and 2009, for durations between four months and two years.

In spite of the fine words spoken at the GDS, in spite of the comprehensiveness and sophistication of the agreement, and in spite of the really significant major learnership and EPWP projects, the GDS remains an unfinished project. Most of the detailed decisions were not implemented a year later, and some of the really big issues around employment have not yet even been discussed. Perhaps it was frustration regarding the impact of the GDS that led President Mbeki to structure the ANC's 2004 election campaign around the theme of confronting poverty and unemployment.

Notes

1 This argument is extended in Brian Levy's 1992 paper, 'How can South African manufacturing efficiently create employment?'.

2 The poverty line used was computed by Haroon Bhorat for a 2003 paper prepared for The Presidency's Ten Year Cabinet Review. It indicates an annual household income of R12 982.50 as the poverty line level.

Document(s) 7 of 11