Unemployment and aids: the social‐democratic challenge for South Africa

Abstract

Addressing Aids and unemployment is one of the greatest challenges facing South Africa at present. Health and welfare interventions will be costly and are likely to meet with resistance from taxpayers. Expanding employment (which will help alleviate poverty and expand the pool of taxpayers) is thus necessary. An inclusive social accord could help, but only if organised labour is prepared to make concessions, as was the case in the new social accords in Australia, Ireland and the Netherlands.

1 INTRODUCTION

There are two major economic and social security challenges facing South Africa: addressing large‐scale unemployment and the Aids pandemic. As of 2003, an estimated 14 per cent of all South Africans were HIV‐positive. According to the government household and labour force surveys conducted from the mid‐1990s onwards, about a third of the labour force is without work (Nattrass, 2000a). This amounts to about 4,7 million people and without question constitutes a socio‐economic crisis of major proportions. The life chances and living standards of entire households are compromised when working‐age adults cannot find employment (Seekings, 2003b). Households burdened by Aids are in an especially difficult position (Desmond et al., 2000; Booysen, 2002; Booysen et al., 2002; Steinberg et al., 2002a, 2002b).

Addressing Aids and unemployment poses major challenges for social solidarity in South Africa. Over the past decade, the labour market and industrial policy environment benefited firms and sectors with relatively high productivity (Nattrass, 2001). Business thus had strong incentives to reduce its dependence on unskilled labour. Once the price of highly active antiretroviral therapy (HAART) started to fall from 2001 onwards, business could supply it, either directly or indirectly through medical aids, to their increasingly skilled workforce (Nattrass, 2003). Rosen & Simon (2002) discuss how business is shifting the burden of Aids onto households and the state, while Rosen et al. (2000) consider when business has an incentive to provide Aids treatment to its workers. Those people without jobs have neither access to earned income nor life‐prolonging medication.

In August 2003, the government signalled its support in principle for the provision of HAART in the public sector. Many unemployed people with Aids will thus be able to access treatment, although this will depend on the scale and pace of the roll‐out. A full‐scale treatment intervention that reaches all who need it is feasible, but will require a substantial commitment of resources (Geffen et al., 2003). If resources are not directed from other priorities, the cost burden will fall on income earners in the form of higher taxation. Given South Africa's high levels of unemployment, this means that the burden of providing treatment for all will fall on a relatively small pool of income earners.

Under these conditions, employers and workers may calculate that they stand to benefit more from a more limited (and less expensive in terms of increased taxation) public sector treatment intervention than a programme providing universal access. Two of the three leading South African macroeconomic models predict that the pandemic will increase per capita income because the impact will be greater on population numbers than on economic growth (Nattrass, 2003). If the Aids pandemic is perceived as being likely to result in an increase in per capita income, then the elite may consider it in their best interests to do very little of significance to halt the epidemic or alleviate its consequences. Those with the economic means to protect themselves and their families against HIV infection (e.g. by providing access to education, condoms, a healthy diet and safer lifestyle choices), and who have access to medical schemes to treat themselves and their loved ones if they become infected, may think their interests are better served by a scenario in which very little is done. They may privately calculate that they stand to benefit more as individuals from a set of policies that prioritise economic growth and minimise taxation than they would from a social response that includes universal access to HAART and entails higher taxation and spending cuts in other areas. They would, of course, be wrong to think that they can entirely insulate themselves in this way from the Aids pandemic; but if they believe they can, this course of action may seem preferable.

This situation has implications for social solidarity regarding Aids treatment. For example, organised labour may well balk at the tax implications of a full‐scale, tax‐financed programme of Aids intervention. Many workers are already able to access HAART through their employers or medical aids. Most of them live in urban areas, which are at the front of the queue in the treatment roll‐out because the greatest capacity to deliver treatment is found in the large urban hospitals. Employed workers may thus have an incentive to support a limited roll‐out with correspondingly less onerous tax implications for their pay packets, rather than a large‐scale intervention aimed at reaching all affected people.

The underlying structural problem is South Africa's high unemployment rate, especially among the less skilled. Section 2 of this article places South Africa in a comparative perspective and summarises the historical roots of the unemployment crisis. Section 3 discusses various ways of addressing the unemployment problem in light of the Aids pandemic, while Section 4 considers the question of how to combat Aids and unemployment/poverty through a social accord process.

2 South Africa'S UNEMPLOYMENT AND WELFARE CRISIS

Until the mid‐1970s, the South African economy was plagued by chronic labour shortages. In this regard, the country experienced the typical sub‐Saharan problem of labour‐constrained development: either a labour supply had to be created through extra‐economic coercive mechanisms, or relatively high wages had to be offered in the capitalist sector (Karshenas, 2001). The uniquely South African solution to this problem comprised a set of racially discriminatory and coercive policies that undermined independent peasant production and channelled relatively cheap African labour to mines, commercial farms and industry. Apartheid policies affected different parts of the economy in different ways (Lipton, 1985) and probably eroded long‐term growth (Moll, 1991). Nevertheless, the coercive creation of a cheap labour force was a crucial ingredient in the early phase of South African economic history.

Not only was the process of de‐agrarianisation and proletarianisation more extensive in South Africa than elsewhere in sub‐Saharan Africa, but the pace of industrial development also far outstripped that of the rest of the continent. Development, fuelled in part by gold revenues and foreign capital, sucked labour from traditional agriculture and facilitated rapid urbanisation. This engine of growth slowed down sharply in the mid‐1970s – average economic growth dropped from 5,6 per cent per annum in the 1960s, to 3 per cent in the 1970s, and then to about 1,5 per cent in the 1980s. Apart from the brief gold‐financed boom in the early 1980s, the South African economy has performed poorly ever since. This has had serious consequences for average living standards. Real per capita income in 2002 was lower than it was 20 years earlier.

By the end of the 1970s, open unemployment had become a major problem. The situation worsened as the economy limped through the late 1980s and into the 1990s. At the dawn of democracy in 1994, over a third of the African labour force was unable to find work (Nattrass, 2000a). Conditions declined further in the 1990s as formal sector employment contracted significantly. As can be seen in Figure 1, manufacturing employment increased marginally in the mid‐1990s, but dropped sharply thereafter. Public sector employment helped boost overall employment until the mid‐1990s, but since 1996, it has contracted alongside private sector non‐agricultural employment (Figure 1). Agriculture has also shed significant numbers of jobs (Simbi & Aliber, 2000). Informal employment rose slightly in the late 1990s, but not sufficiently to prevent the trend rise in unemployment.

Trends in non‐agricultural formal employment in the 1990s

Unemployment rates are especially high amongst the youth (Nattrass, 2002) – the group most devastated by Aids. Figure 2 shows that South Africa's unemployment rates, when compared with other middle‐income developing countries, are literally ‘off the charts’. This is a result of South Africa's comparatively weak historical performance with regard to both employment and output growth (see Figures 3 and 4). To make matters worse, South Africa's growth path has become even less employment friendly since 1990, largely because of ongoing structural change, the impact of technological change and rising wage and non‐wage labour costs (Fedderke & Mariotti, 2002).

Unemployment rates in middle‐income countries

Index of employment growth in Ireland, Australia, the Netherlands, the European Union, middle‐income countries and South Africa

Index of output growth in Ireland, Australia, the Netherlands, the European Union, middle‐income countries and South Africa

A range of factors has contributed to South Africa's poor employment performance since the mid‐1970s. These include slow growth, tariff protection for increasingly capital‐intensive sectors and large‐scale strategic investments by the state (Kaplinsky, 1995). During the 1970s and early 1980s, the coincidence of rising wages and negative real interest rates meant that the cost of capital relative to labour fell to about half the level it had been in the 1960s (Meintjes, 1998: 11–12). Tax breaks for capital investment further encouraged firms throughout the economy to adopt more labour‐saving techniques. The change to positive real interest policies and the depreciation of the rand in the mid‐1980s reversed the downward trend in the user cost of capital but failed to boost employment growth significantly. The net result was that the South African industrial sector became steadily more capital‐intensive over time.

One of the contributing factors to this trend has been developments in the labour market. Under apartheid, independent African trade unions were denied access to industrial councils, where wage bargaining took place at industry level. In 1979, the Wiehahn Commission recommended full trade union rights for Africans. This was implemented shortly thereafter and, by the mid‐1980s, African trade unions were participating in industrial councils (now called ‘bargaining councils’). Not all workers, however, were covered by industrial council agreements. The impact of deracialising the industrial council system was thus to ‘recycle’ the old apartheid wedge between white and black workers into a wedge between ‘insiders’ (i.e. workers covered by industrial councils) and ‘outsiders’ (i.e. those in poorly paying, uncovered sectors such as agriculture and services, and the unemployed) (Moll, 1996).

African trade unions were able to use their new‐found institutional muscle to push up wages, especially for relatively low‐paid workers (Hofmeyr, 1994). To the extent that such wage pressures encouraged firms to adopt labour‐saving techniques (Nattrass, 2000b), these trends would have contributed to the increase in capital intensity. Most calculations of employment elasticity in South Africa indicate that the labour demand curve is relatively elastic. Estimates typically range from –0,66 to –0,85, which suggests that a 10 per cent increase in wages will result in a drop in employment of between 6,6 and 8,5 per cent (Eager Report, 1999: 7). Other pressures include the rise in non‐wage labour costs, such as costs relating to hiring and firing and, of course, costs relating to Aids (Nattrass, 2003).

2.1 Wages, employment and profitability, 1990–2001

Those who lost or failed to find formal employment were the big losers in the 1990s. By contrast, the owners of capital and those (predominantly skilled) workers who retained their jobs, did relatively well. Figure 5 shows that real average remuneration rose as employment fell during the 1990s. As the South African economy restructured towards the more skill‐ and capital‐intensive sectors and enterprises, the skill composition of employment shifted in favour of more skilled (and better paid) workers (Fedderke & Mariotti, 2002: 840–1). Formal employment now has a far higher concentration of ‘good jobs’, although there have generally been fewer jobs for others.

Index of labour productivity, employment, average wages and profitability in South Africa

How has this trend affected profitability? On one hand, firms have had to contend with sharp increases in labour costs. On the other hand, however, they have been able to secure an increase in labour productivity. By restructuring and downsizing their workforces, they have ensured that each remaining worker contributes more on average to output than had been the case when employment was higher. As they have been able to secure faster growth in productivity than real average wages, this has been advantageous for profitability.

One rough indication of profitability is the gross profit share (i.e. the share of gross output accruing to the owners of capital). Broadly speaking, if the growth in labour productivity is greater than the growth in real wages, then workers are contributing more to output growth than they are getting back in wages, and hence the share of output going to capitalists (the profit share) will rise. Figure 5 shows that the average rate of growth of productivity exceeded that of real wages for most of the 1990s. As a result, the aggregate profit share was about 10 per cent higher in 2001 than it was in 1990. Capitalists have also benefited from rising rates of profit (i.e. the rate of return on capital) in most sectors. This can be seen in Figure 5, which shows how the average net profit rate has increased since 1990. This is the average for all sectors, excluding community social and personal services, which comprise most government services.

The general rise in labour productivity and the proportion of skilled workers in total employment in the 1990s is consistent with government policy to drive the economy ‘up the value chain’, i.e. to force the economic structure to shift towards a more skill‐intensive growth path. The high‐productivity strategy entails a mixture of incentives to encourage training and the development of high value‐added forms of economic activity, as well as continued support for those aspects of labour‐market policy that hinder low‐wage, labour‐intensive job creation (Nattrass, 2001). As argued above, the strategy appears to have benefited those workers who managed to keep their jobs because real wages grew significantly. However, unlike other countries such as Ireland that have successfully pursued it, the strategy has done nothing yet to improve the ability of the South African economy to create jobs.

South Africa's growth strategy may well deliver benefits in terms of income growth in the longer term. Its success, however, is dependent on skills development and on high and sustained rates of investment. The government's orthodox economic policy – the Growth, Employment and Redistribution (GEAR) framework – is premised on investment becoming the driving force for growth (Nattrass, 1996). The hope was that investors would respond well to falling inflation and the budget deficit. Unfortunately, investment has not responded as quickly or extensively as hoped by the GEAR macroeconomic modellers, and growth has been sluggish for the past decade. Some of the reasons for this lacklustre performance were beyond the control of economic policy makers. The Asian crisis and overzealous monetary policies by the independent Reserve Bank both acted as unexpected economic brakes. It is a moot point whether the government should have continued with its restrictive fiscal policies given the recessionary conditions of the time (Weeks, 1999).

Rapid capital accumulation (and even rising capital intensity) is desirable if it is accompanied by significant growth in employment. The policy gamble taken by the South African government is that the shift to greater capital‐ and skill‐intensity will provide a new engine for sustainable growth in the medium to longer term. Unfortunately, skill shortages (which are currently driving up the price of skilled labour relative to unskilled labour) will continue to act as a constraint on growth, and the brain drain of young professionals will exacerbate the situation for some time. The Aids pandemic is likely to worsen the skills shortage (Nattrass, 2003).

In short, the South African growth path has become increasingly skill‐ and capital ‐intensive and has delivered benefits to both capital and labour. In this post‐apartheid ‘distributional regime’, the unemployed have been the biggest losers (Seekings & Nattrass, 2004). They are marginal to South Africa's high productivity growth path, and their deaths from Aids are likely to have less of an impact on the economy than the deaths of employed workers. If the only objective of government policy was to promote growth, then prolonging the lives of the unemployed would probably feature low down the government's list of priorities (Nattrass, 2004). This, of course, could be a miscalculation because providing treatment for those with Aids (whether employed or unemployed) is likely to lower the rate of HIV transmission (Johnson & Dorrington, 2002). In other words, the social benefits exceed the sum of the individual benefits for HAART patients.

Ultimately, the discussion about Aids interventions has to entail social concerns and values. Should we as a society sit back and allow all those who are marginal to the process of income generation die – or should we do something about it? What is the reasonable response to this major challenge for social justice? The problem with this question, however, is that it immediately poses the question of what else we could be doing as a society to help the unemployed. What is an appropriate balance between addressing Aids and addressing unemployment and poverty? Given the strong connection between poverty and HIV transmission (Stillwaggon, 2002; Nattrass, 2004), it follows that antipoverty measures should be an integral aspect of any Aids intervention strategy. The challenge, in other words, amounts to how best to address poverty and Aids simultaneously. This issue is addressed in Section 3 below.

3 ADDRESSING UNEMPLOYMENT, POVERTY AND AIDS

There are two broad ways in which the government could respond to the challenge of unemployment and poverty: through direct interventions (e.g. income support and/or public works) and through policy measures to make the economic growth path more labour demanding.

3.1 Direct government intervention

South Africa's welfare system provides for the young (through child grants), the elderly (through generous non‐contributory pensions) and the disabled. The underlying assumption is one of full employment, i.e. that able‐bodied adults can provide for themselves through work (Nattrass & Seekings, 1997). As a result, unemployed people rely on pensioners, transfers from employed family members and (increasingly) on disability grants to survive.

The South African government faces significant fiscal exposure to the Aids epidemic through the welfare system. A government means‐tested disability grant of a maximum of R700 a month is available to all ‘severely physically and mentally disabled people’ older than 18. This includes people living with Aids. However, there is no clear policy on Aids‐related disability grants (Boulle, 2003). Means‐tested caretaker grants are available to the caretakers of disabled children, including those affected by Aids.

In 2001, 643 000 people were receiving disability grants which, in the opinion of Van der Berg & Bredenkamp (2002: 50), was a surprisingly low figure given the eligibility criteria. However, by October 2002 the number of disability grants had risen sharply to 831 271. Simkins (2003: 8) observes that this ‘probably reflects an increase in take up rates as well as the rising number of people disabled by the development of full‐blown Aids’. Between April 2001 and June 2002, the number of child‐care dependency grants also increased substantially. This trend may partly be the result of ‘the emergence of claims for children seriously ill as a result of HIV infection’ (Simkins, 2003: 11). Assuming a constant take‐up rate for the non‐Aids disabled, and a rise in the take‐up rate for the Aids‐disabled, Simkins estimates that the number of disability grants will rise to 1 236 847 in 2010. The Fiscal and Financial Commission has warned that this figure may be an underestimate, and that more Aids‐related disability grants are likely in the future (Cape Times, 2003a). The costs to the government will rise from R6,4 billion to R9,5 billion as a result (Simkins, 2003:9). Costs associated with care dependency grants could also increase substantially due to Aids (Barberton, 2000).

The government, after years of procrastination and obfuscation, finally agreed to make HAART available through the public sector, albeit on a limited and highly controlled basis. Those who are able to access the programme and who respond well to treatment will in many cases no longer be eligible for a disability grant – although this seems to vary across the country. The government would thus ‘save’ a significant portion of increase projected by Simkins. The problem with this scenario, however, is that the loss of the disability grant may have serious financial and other implications for people living with Aids. Welfare transfers, like pensions and disability grants, are important components of household income – particularly for the very poor (Leibbrandt et al., 2000; Seekings, 2003b).

For many people living with Aids the disability grant is, unsurprisingly, a major source of relief. One of the respondents interviewed as part of a broader study of the impact of Aids, went as far as to say ‘I love this HIV’ because of it. She explained her choice of words as follows (Steinberg et al., 2002b: 29):

Yes, I like this HIV/Aids because we have grants to support us… Before I was staying with my mother and father and sister, they didn't work. Maybe I was taking three to four days without food. People discriminated against me and no one came in the house. The only thing that was helping was my grandmother's pension. We were surviving on that money. Concerning the illness, our lives are changed completely.

The notion that someone might ‘love this HIV’ seems shocking; but it is understandable when one considers the desperate circumstances in which households can find themselves when they lack an income earner. The advent of a disability grant, as was clearly the case for the respondent quoted above, is a major lifeline for the entire family. The threat of its removal as a result of HAART is serious indeed. There is thus a case for reducing, but not removing altogether, the disability grant for those on HAART. This, however, raises a further question: why privilege people on HAART (whose immune systems have been reconstituted as a result) over those who are equally poor, but are HIV‐negative? In other words, rather than introducing a new, lower grant to replace the disability grant, why not introduce a basic income grant (BIG) for all South Africans?

Welfare support for the poor could be provided in two ways: via a means‐tested system (a ‘dole’) or through a universal BIG. The advantage of the latter is that it wastes fewer resources on bureaucracy. A BIG of R100 per month for all South Africans could contribute substantially to reducing poverty and inequality in South Africa. According to Bhorat (2002), the numbers of people living below the poverty line would fall by about two‐thirds. Interestingly, there appears to be widespread support for the introduc ‐tion of a BIG in South Africa, ranging from the opposition party, the Democratic Alliance, to the socialist left (Matisonn & Seekings, 2002). However, there is less agreement about how to finance such a grant. The recent Taylor Committee report on comprehensive welfare reform argued in favour of a BIG (Taylor Committee, 2002), but was curiously silent about how it should be financed.

The People's Budget (www.cosatu.org.za) is supported by the Congress of South African Trade Unions (Cosatu), the South African Council of Churches and the South African NGO Coalition (Sangoco). It proposes that part of the needed revenue could be raised through a ‘solidarity levy’ in the form of a 17,5 per cent surcharge on income tax for the top two quintiles – and the rest in the form of increased taxation of ‘the high‐income group’. This is broadly in line with the resolution of Cosatu's 7th National Congress that the cost of the BIG must ‘fall on the rich’. By contrast, Le Roux (2002) proposes that the BIG be financed by a 7,3 percentage point increase in value‐added tax (VAT) and a 50 per cent increase in excise and fuel taxes. The advantage of Le Roux's proposal is that it is broad‐based and redistributive. Those who spend more than R1 000 a month end up paying more in consumption taxes than they benefit from the R100 universal grant.

Despite the wide‐ranging support for a BIG, the South African government has been reluctant to endorse it. One source of concern is the ‘hand‐out’ nature of the grant, which in many circles (including the government) is deemed less desirable than providing people with the ‘dignity of work’ through direct job creation. Those who believe that citizens have an obligation to work are also opposed to a BIG. Another concern, perhaps, is that the R100 a month, once implemented, could become a political site of struggle over the size of the grant. As Le Roux (2002) has shown, the net cost of a BIG (financed out of VAT) will be about R15 billion a year, i.e. a few billion more than the government is currently spending on the gross cost of old‐age pensions. If the grant were to double or triple (as a result of demands from voters), then the cost implications would be serious indeed.

There are, of course, other ways of channelling income into the hands of the unemployed. An obvious alternative to a grant is to ‘self‐target’ the unemployed by offering low‐wage jobs through a government‐funded national public works programme (PWP). Low‐wage PWPs target the poor (because only the poor will work for low wages) without undermining the labour market. Properly designed PWPs have the potential to alleviate poverty (McCord, 2002) and have the additional benefit of creating assets (e.g. building roads, removing alien vegetation from watercourses, etc.). The disadvantage of PWPs relative to a BIG is that a substantial proportion of resources (typically between 40–50 per cent) are absorbed by administration, rather than being channelled more directly to the poor via wages. The more complex the PWP (e.g. those that attempt to provide some skills training as well), the greater the proportion of resources absorbed by management, and the greater the risk of inefficiency and failure.

Nevertheless, as the experience of Chile has shown, it is possible to absorb a high proportion of the unemployed in government‐funded and managed PWPs (Cortaza, 1997). In 1997, 5,5 per cent of the labour force was employed in the Chilean government's Minimum Employment Plan at wages equal to one‐third of the minimum wage. This rose to 8,5 per cent in 1982 (Cortaza, 1997: 237–9). According to McCord (2003), it would cost South Africa R22,8 billion to provide low‐wage employment for 2,6 million unemployed people a year (i.e. about one‐third of the 7,7 million who report that they do not have jobs and would like one). This assumes that each person works ten days a month for R35 a day and that the wage bill absorbs 48 per cent of the total resources required to run a national PWP. (The 48 per cent wage bill figure was based on the average for the National Economic Forum job creation programme between 1992 and 1998.) This is more than the net cost of a BIG which, as noted above, Le Roux estimates to be in the region of about R15 billion a year, and about the same as the (mid‐point estimate of) the cost of a full‐scale Aids intervention including HAART (Nattrass & Geffen, 2003).

Whether the South African state has the capacity at all levels of government to implement such a national PWP successfully is a good question. Recent scandals concerning corruption and inefficiency in the delivery of pensions in the Eastern Cape indicate that in poorer parts of the country such capacity is likely to be lacking. This is a major problem from the point of view of inequality because it is precisely these areas that require well‐targeted and well‐designed PWPs the most. A decided limitation of PWPs is the high proportion of expenditure that is channelled to the bureaucrats administering the projects rather than to the wages of beneficiaries. If the bureaucrats do not carry out a good job in ensuring that the PWPs reach the poorest people, and in the most efficient manner, then the redistributive impact of the programme is seriously weakened. Under the circumstances, then, it is likely that a BIG will be better targeted than a national PWP and is likely to reach more of the poor, and also in a more effective manner. A BIG has the added advantage in the context of the Aids pandemic of reaching those who are too ill to participate in a PWP.

Given that income earners are beneficiaries of the current economic growth path, an appropriate social‐democratic response may well be higher taxation in order to finance a BIG. Higher taxation was part of the implicit social contract embedded in the Scandinavian social‐democratic model. Such ‘social accords’ sought to balance the need to maintain profitability (so as to encourage investment and ensure sustainable growth), while ensuring a reasonable growth in wage income and the provision of basic income support and training for the unemployed (Nattrass, 1999).

It has been estimated that implementing a full‐scale Aids prevention and treatment intervention that provides HAART to all those who need it would require an increase in resources equivalent to raising VAT by between 3 and 7 percentage points, depending on what level of care is provided to those suffering from Aids‐related illness (Nattrass & Geffen, 2003). (The money could, of course, be raised through income tax rather than VAT. The discussion about taxation is presented here in terms of VAT simply to keep the argument simple.) It is, however, important to remember that these costs assume a rapid and extensive take‐up of HAART, and no rationing of Aids patients in the public sector. This is why this particular estimate is higher than the other available costing studies that assume greater rationing and more limited treatment regimens (see Boulle et al., 2003). The estimate was explicitly designed to answer the question of what resources were needed to provide treatment to all who need it. If the government decides to provide only a limited and rationed treatment programme, then total costs will be correspondingly lower.

According to Le Roux's (2002) estimates, an increase in VAT of 7 per cent is required to finance a BIG. Le Roux's (2002) proposed increase in VAT to fund a programme that costs R15 billion is larger than the VAT increase required to fund a PWP costing R22,8 billion. This is because the direct (or gross) costs of a BIG (R52 billion) are substantially larger than the net costs. By financing the BIG out of an increase in VAT, the tax system will ‘claw back’ the grant from richer people because they will end up paying more in expenditure taxes than they benefit from the grant. Le Roux also assumes increases in excise taxes. In other words, if both a BIG and a full‐scale Aids intervention programme were to be implemented, this would require a significant increase in taxation. Assuming the cost estimate reported in Nattrass & Geffen (2003), the required tax increase would be in the region of increasing VAT by 10–14 percentage points. The question is whether this is feasible.

There is no exact technical answer to this question, as different societies tolerate different levels of taxation, and at different times. Welfare expenditure as a proportion of gross domestic product (GDP) has risen with economic development and in times of crisis (such as war) citizens have accepted large increases in taxation as legitimate (Seekings, 2003a). The notion of what is and is not ‘affordable’ thus varies according to the social and economic context. Given the scale of the unemployment problem and the Aids epidemic, it is possible (but probably unlikely) that reasonable South Africans might agree to an increase in taxation so as deal with it. While this proposal would run counter to the government's orthodox fiscal policies, there are some grounds for believing that the South African economy may be able to cope with an increase in the tax take – especially given that the negative impact of increased taxation on demand will be ameliorated in part by the injection of demand into the hands of the poor through the BIG.

According to one (albeit rather optimistic) view, South Africa's average tax rate is ‘significantly less than that which would be predicted given the country’s economic profile' and that ‘an additional R25 billion per year could be mobilised without undermining international competitiveness’ (Samson, 2002: 91). Assuming for the sake of the argument that this assessment is correct, an additional R25 billion could go a long way towards financing both a BIG and a large‐scale Aids intervention programme. Taking Le Roux's estimate of R15 billion a year for a BIG, and assuming the lower‐bound cost estimate of R14,1 billion for a full‐scale Aids intervention programme (Nattrass & Geffen, 2003), we would need to find about R4 billion a year in addition to the resources mobilised by this proposed increase in taxation. Some relief could be sought by applying for grants from the Global Fund for Aids, Tuberculosis and Malaria, and from other international aid agencies. This was the route followed by Botswana.

Another option is to drive down the costs of the Aids treatment programme through an aggressive approach to drug pricing. Engaging in Brazilian‐style negotiations with pharmaceutical companies over the price of antiretrovirals would help, as would providing further support for the domestic production of generic medication. Brazil has made a concerted effort to produce generic antiretrovirals domestically. Between 1999 and 2001, the proportion of antiretrovirals produced in Brazil rose from 47 per cent (19 per cent of government expenditure on antiretrovirals) to 63 per cent (43 per cent of expenditure). The remainder is purchased on the international market, and Brazil has had great success in negotiating price reductions (Galvao, 2002: 1864). Brazil has threatened to engage in compulsory licensing (i.e. breaking patents) in order to force price concessions from pharmaceutical companies. For example, after Brazil threatened to break the patent on Efavirenz, Merck agreed to a 60 per cent price reduction. Merely the threat to do so has proved sufficient to bring about a significant reduction in costs. The decision in November 2001 by the World Trade Organisation to allow the use of compulsory licensing in cases of national public health emergencies has further strengthened Brazil's hand – as well as that of other middle‐income developing countries (such as South Africa) that have the industrial capability of producing antiretrovirals if necessary.

Even if such measures were introduced, it is important that the range of policy options needs to be broadened beyond the size and allocation of the government budget to address the root cause of the poverty problem in South Africa: high unemployment. Ideally the government should be proceeding on several fronts, such as implementing a BIG and Aids interventions as well as encouraging the expansion of more labour‐intensive firms and sectors. This means addressing labour‐market policy and the economic growth path itself.

3.2 Unemployment and social accords: lessons from international experience

As argued in Section 2, the South African growth path has become steadily less labour‐absorbing. The high‐productivity growth strategy has helped foster greater skill and capital intensity, but it has done little to create jobs for the many (relatively unskilled) people who need them. Under the post‐apartheid distributional regime, labour productivity has grown along with real average remuneration and profitability. Those persons with jobs have done relatively well; those without have born the brunt of rising poverty. The question is whether anything can be done to make the growth path more receptive to generating employment.

Judging from the international experience of labour‐demanding development, the following factors seem to help: an export orientation; wage growth slower than that of productivity; flexible labour markets; efficient and well‐targeted state support for key economic sectors; and development‐oriented macroeconomic policies (Nattrass, 2001). However, pointing to past experience is not always helpful because the economic context may be different today than it was in the past. For example, the present trade regime rules out a variety of government interventions that were used to good effect in the past by newly industrialising economies such as South Korea. Moreover, the particular social and economic context that facilitated a set of labour‐market policies, such as those found in South‐East Asia, does not exist in South Africa.

More specifically, South Africa has an institutional commitment to tripartite forms of negotiation, as evidenced most clearly in the National Economic Development and Labour Council (Nedlac), South Africa's peak‐level tripartite institution, and to encouraging cooperation between labour and capital at industry and firm level. This sets South Africa apart from other African countries and most other middle‐income economies. A more useful approach, then, is to ask what lessons can be learned from the recent experience of countries that have experienced strong employment growth in a context of consultation and cooperation between labour, capital and the state. The experience of Australia, Ireland and the Netherlands is instructive in this regard. Like South Africa, these countries experienced an employment crisis, and had strong trade unions and a tradition of collective bargaining and tripartite negotiation. In each case, the trade union movement made significant concessions in order to restore profitability and employment growth.

Broadly speaking, employment can rise if growth increases rapidly, or if the growth path becomes more labour intensive. If employment is driven by rapid growth, then average labour productivity and employment could rise. The best recent example of this high‐productivity/high‐output/high‐employment growth path is that of Ireland. Figure 3 illustrates how rapidly employment grew in Ireland. As output growth was even more spectacular (Figure 4), average labour productivity (i.e. output per worker) also rose significantly (Figure 6). If, however, the main driver of employment growth is the expansion of part‐time or low‐productivity jobs, then employment could rise faster than output, thus resulting in a decline in average labour productivity. This was the growth path experienced by the Netherlands in the 1980s (see Figures 3, 4 and 6). The Australian experience represents something of a middle path between the two.

Index of labour productivity in Ireland, Australia, the Netherlands, the European Union, middle‐income countries and South Africa

The roots of the Australian Accord can be traced to 1982, when a mini‐resources boom sparked off by the increase in the gold price led to excessive wage increases, inflation, balance of payments problems and a decline in output and employment. According to the Research Officer of the Australian Council of Trade Unions (Harcourt, 1995):

[This] prompted the union movement to collectively take responsibility for putting employment first when framing economic strategy… [and to take]… responsibility for ensuring Australia's international competitiveness in terms of low inflation and falling real unit labour costs.

The initial phase of the accord entailed wage restraint in return for lower taxation and improved welfare programmes. However, as growth remained sluggish, all parties agreed to real wage cuts and declining government expenditure. This was accompanied by a steady decentralisation of wage bargaining to firm level, and greater concentration on improving productivity through workplace and industry‐level initiatives (Nattrass, 1999). Trade union support for this dramatic shift in strategy was the key factor. As can be seen in Figures 3, 4 and 6, Australian output and employment recovered to grow strongly (along with labour productivity) in the 1990s.

Ireland and the Netherlands (as well as Italy, Portugal and Spain) are quintessential examples of the ‘new social pacts’ that have arisen recently on ‘apparently arid ground’ in Europe (Rhodes, 2001: 167). Both have, to varying degrees, entailed the fashioning of innovative links between labour‐market policy (pertaining to training, employment conditions and wage bargaining), industrial policy, welfare policy and taxation. In each case, the trade union movement made real concessions with regard to wage restraint and labour regulation in return for benefits such as lower taxation, enhanced representation, skills development and other productivity‐enhancing policies and measures.

Figure 3 shows that recent employment growth in Ireland and the Netherlands compares favourably with that of the European Union as a whole. As was the case in Australia, an employment crisis preceded each social accord process. A commonly perceived crisis, and recognition amongst all parties that concerted action is required to address it, seems to be a precondition for a social accord to be reached.

As noted earlier, there is a striking difference between the Dutch and Irish experiences with regard to productivity growth. Rapid rates of growth underpinned rising productivity in Ireland, whereas the expansion of jobs (most notably through the provision of part‐time work for women) in the Netherlands resulted in falling, and then slow, productivity growth. Lower average labour productivity is to be expected when employment growth is promoted in this way through part‐time employment (which amounts to a form of job sharing) and through government‐sponsored job creation schemes. Productivity enhancement at firm level is crucial for maintaining growth in profits and wages. However, if raising productivity becomes the sole policy objective, then there is a danger that the growth path will exclude the many unemployed who wish to work. The Dutch solution to this problem was to ensure that productivity‐enhancing measures (such as training) were accompanied by measures to boost employment, even though this implied lower average labour productivity.

The Irish social accord process is the most wide‐ranging of the new social accords. It began as an attempt to restore profitability, investment and growth, and subsequently expanded in scope to include issues of education, health, social welfare and incomes. The first neocorporatist accord was struck in the mid‐1980s, when Ireland was mired in slow growth, rising public sector deficits and high unemployment. Subsequent pacts negotiated in 1990, 1993, 1996–7 and 2000 linked income policies to wage restraint and reforms in taxation, education, health and social welfare. Each agreement was preceded by the publication of a discussion document prepared by the National Economic and Social Council – an advisory body composed of representatives of the social partners and senior government officials. As the pacts developed over time, more emphasis was placed on supply‐side measures to promote training and productivity growth. Over time, the accord process was expanded to include more representatives of civil society.

As is the case in Australia, Irish labour market and industrial strategy is geared towards ‘best practice’ and productivity enhancement. This has resulted in most of the agreements being ‘largely tailored to the demands of the insider unionised sector and… protecting the post‐tax income of the employed insiders’ (Kavanagh et al., 1997). ‘Partnership 2000’ was negotiated with a wider grouping of social partners, but maintained a strong focus on improving industrial democracy and training in an effort to boost productivity through best practice and consensual labour relations.

Rapid growth of investment is at the heart of the Irish success story. Investors are interested in profitability and stability/predictability in the business environment – the Irish accords delivered both. Wage restraint in the late 1980s and early 1990s helped ensure that profitability doubled between 1987 and 1996 (Lane, 1998). As O'Donnel (2001: 11) puts it, ‘the resulting environment of wage moderation and high profitability is almost certainly a key factor in Ireland’s employment creation, attraction of inward investment and the unprecedented commercial success of indigenous companies'.

The experience of Ireland shows that an inclusive, cooperative growth path underpinned by productivity enhancement is possible, but that this has to be based on strong growth. Ultimately, this means ensuring that the economic environment remains conducive to investment. The issue that concerns all social accords (both old and new) – i.e. ensuring that wages do not squeeze profitability too far – is always lurking on the horizon. When growth slows down, this becomes the central issue.

4 A SOCIAL ACCORD IN South Africa?

Social accords arise from a common sense of crisis and recognition of the need to restore profitability in order to ensure the sustainable growth of employment and income for all. As shown in Figure 5, both the rate and share of profit in South Africa rose for most of the 1990s – so it appears that profitability is being restored in the absence of a social accord. Rather than bargaining with labour over their strategy, South African firms are simply responding to the economic and policy environment by shedding unskilled labour and by ensuring that wage growth is matched by improvements in productivity. They have, in short, succeeded in recreating the conditions for renewed accumulation without an explicit commitment to wage restraint on the part of organised labour.

If the only objective of a social accord is to restore profitability then South Africa does not, on the face of it, appear to need one. If, however, the objective is to facilitate a stronger, more labour‐demanding and less conflictual growth path, then a social accord process could potentially be of value. An explicit agreement on the part of organised labour to restrain wage demands in line with productivity growth could help improve the investment climate (as was the case in Australia, Ireland and the Netherlands). There certainly is room still for a more cooperative growth path that could benefit both labour and capital.

Such an accord could include the following:

1.	A framework agreement (probably negotiated in Nedlac) detailing agreed parameters for a wage increase. These could be blanket wage increases for all sectors (as in the Irish social accord), or could be stratified by sector.
2.	Industry‐level wage bargains would be constrained by the framework agreement and procedures put in place (as in the Irish case) to accommodate those firms that are unable to afford the wage increases.
3.	The government continues to provide support for training and skills development through its various dedicated industrial and labour policies.

While this kind of ‘insider’ accord could help improve growth at the margin, it is unlikely to have much impact on employment in the short or medium term. As shown by the Irish (and, to a lesser extent, the Australian) social accord processes, the only way for a productivity‐based accord to generate significant benefits to employment is if growth is tremendously fast and fuelled by domestic and foreign investment. Such a scenario is unlikely in South Africa, given the chronic problem of skills bottlenecks.

If South Africa is to address the employment problem by means of a social accord, then the agreement has to be inclusive of labour‐market ‘outsiders’. Whether this means broadening the parties to the agreement to include the unemployed and civil society organisations (as in the Irish case), or simply mandating the government to look after the interests of the unemployed (as in the Netherlands), remains an open issue.

An inclusive social accord in South Africa would need to support high‐productivity activities, but not at the cost of slower employment growth. Training and skills development should continue, as this will improve the competitiveness of high‐productivity sectors. At the same time, however, employment needs to expand, perhaps in the form of job sharing, or in lower‐wage, labour‐intensive activities and even in government‐sponsored PWPs. This will serve to reduce average labour productivity – but as the Dutch growth path shows, this can be an appropriate outcome if the objective is to address the problem of marginalised outsiders by creating jobs for them.

An ‘outsider‐friendly’, more inclusive social accord in South Africa could include the following as additional aspects to the three outlined above:

1.	Organised labour and business agree to labour market reforms in order to encourage the growth of labour‐intensive firms and sectors (e.g. setting lower or no minimum wages for smaller, more labour‐intensive firms).
2.	Where other labour legislation can be shown to be harmful to employment creation (e.g. rules about retrenchment), then additional labour reforms should be considered.
3.	The government agrees to increase the number of PWPs, to remove all taxes on employment (e.g. payroll taxes) and to make all Aids prevention and treatment expenditure at firm level fully tax deductible.

Unlike the social accords in Australia, Ireland and the Netherlands, which provide tax concessions to workers in return for wage restraint, this is not a significant or sustainable option in South Africa where the pressure to redistribute income through the fiscus is substantial. Indeed, as noted above, if a large‐scale government Aids intervention programme and a BIG were to be implemented, then all income earners would be faced with a major tax increase.

One of the obstacles facing an inclusive social accord in South Africa is whether organised labour is prepared to accept the associated labour‐market reforms and tax implications of a strategy designed to boost employment. If South Africa is unable to make progress on points 4, 5 and 6 – and concentrates instead on high‐wage, high‐productivity jobs only, then all that will remain is an ‘insider’ accord that supports a growth path that has little, if any, chance of reducing unemployment in the foreseeable future.

As highlighted in the above discussion, the social accords in Australia, Ireland and the Netherlands saw organised labour making concessions in terms of wage restraint and labour‐market reforms in return for tax cuts and policies designed to promote skills development and training. In South Africa's case, organised labour has already achieved many of these policy gains. This was helped by Cosatu's alliance with the African National Congress and by the fact‐enlightened policy making by the Ministries of Labour and of Trade and Industry with regard to skills development. More recently, income earners have been handed tax cuts by the Minister of Finance. Organised labour in South Africa has, in other words, achieved many of the direct benefits typically associated with the new social accords – without having to make any concessions. It is an ironic possibility that the South African government may have just missed the opportunity to forge a social accord by handing out many of its bargaining chips for free.

Nevertheless, there is still potential for a social accord process to deliver meaningful changes to South Africa's policy environment and growth path. Simply bringing the social partners together to discuss problems has potential benefits. As O'Donnel (2001: 5) remarks about the Irish social accord process, the benefits extended far beyond bread‐and‐butter negotiations between labour and capital:

Bargaining describes a process in which each party comes with definite preferences and seeks to maximise their gains. While this is a definite part of Irish social partnership, the overall process (including various policy forums) would seem to involve something more. Partnership involves the players in a process of deliberation that has the potential to shape and reshape their understanding, identity and preferences.

South Africa is fortunate in having a strong tradition of social corporatism and a national‐level institution in the form of Nedlac that is capable of facilitating discussion and negotiation between the social partners. The main beneficiaries of Nedlac have been labour and capital, but there is potential for drawing in a wider range of social interests and constituencies into an accord process. This could be done by expanding and empowering Nedlac's Development Chamber, which was set up to facilitate the inclusion of civil society (the ‘community sector’) in certain negotiations.

Aids policy has already been the subject of negotiation in Nedlac. In July 2002, the Treatment Action Campaign (TAC) and Cosatu requested Nedlac to instigate a process to negotiate a national HIV/Aids treatment programme. In September that year, the Nedlac Management Committee established a senior HIV/Aids task team comprising high‐level representatives of the government (including the Director‐General of Health and Labour, and the Chief Director of HIV/Aids policy); representatives of organised labour and business; and a representative from TAC and the Community Sector Convenor of Nedlac. Between September and December, the task team met several times and drafted a framework agreement in favour of an Aids treatment intervention programme. The process, unfortunately, broke down when the government refused to sign the agreement – saying variously that there was no agreement, or that as the government it was not required to sign anything of the sort. Rather than encouraging social reflection and the participation of civil society in policy formation, the government undermined Nedlac and adopted a technocratic approach. The Health Minister, Tshabalala‐Msimang, appeared to regard the agreement as a challenge to her power to decide health policy. She was quoted in a major national newspaper as saying: ‘Why must the government enter into agreements with everyone? Tomorrow I must enter into an agreement with asthma sufferers?’ (Mail and Guardian, 2003). This response was unfortunate, because it forced civil society interest groups (such as TAC) to confront the government through legal action and via high‐profile civil disobedience campaigns.

There is, however, some validity to the Health Minister's complaint – at least in so far as it can be read as a concern about government policy being held hostage to specific interests (Aids sufferers, asthma sufferers, etc.). This, in principle, is a potential problem given that not all those with health needs are as well represented as people living with Aids are represented through TAC. However, unlike asthma, Aids poses unique public health challenges to society. It is thus appropriate to raise this health crisis in a national forum such as Nedlac, but only if a broad cross‐section of civil society is included in the discussions.

Equally important is that the issue of Aids interventions should not be discussed in isolation. Addressing the problem of Aids entails more than a mere shift in health policy; it requires rethinking the way that funding is allocated from the national to provincial levels, a reconsideration of spending and taxation decisions and a discussion of the general growth strategy. It should be an integral part of the social accord process to address unemployment. It is a recipe for disaster if Aids is discussed in one chamber while an economic social accord is discussed in another. If the two issues are separated, then the trade‐offs between spending money on (for instance) PWPs, a BIG or a large‐scale Aids intervention will not be confronted squarely. It could also result in a less‐than‐optimal deal regarding labour‐market reforms.

For example, as a result of reasonable deliberation, all parties (labour, business, civil society and the government) may agree that addressing unemployment and combating Aids through a range of interventions including a full‐scale HAART programme is necessary and that taxation would have to rise to fund it. Faced with this consensus, labour may well opt for more broad‐ranging and extensive reforms than it would otherwise agree to, simply in order to spread the load of the required tax increases across a large number of income earners. Organised labour and civil society may also put more pressure on the government to cut back on unproductive expenditure, such as defence, in order to reduce the burden on taxpayers.

Additional information

Notes on contributors

Nicoli Nattrass

Professor, School of Economics, University of Cape Town, Cape Town, South Africa.

Notes

Professor, School of Economics, University of Cape Town, Cape Town, South Africa.