+Bioline International Official Site (site up-dated regularly)

African Population Studies/Etude de la Population Africaine, Vol. 19, No. 2, Sup. A, 2004, pp.47-84

Overview of Poverty Situation and Reduction in South Africa for the Past 10 Years

Director: Population Developmental Research
Chief Directorate: Population and Development
Department Of Social Development
Republic Of South Africa

Poverty and inequality are some of the major developmental challenges that are facing South Africa after ten years of democracy. Since 1994, major efforts have been made by the new government to tackle issues with regard to poverty and inequality within the framework of the Growth, Employment Strategy and the Reconstruction and Development programme within the framework of the South African population policy, which is based on the Cairo declaration of 1994. Strategies that the current government has undertaken to address development amongst the disadvantaged are land distribution, Small, Micro and Medium Enterprises (SMME’s) and the public works programme. However, measures like the Gini Coefficient still indicate that huge inequalities exist among the four racial groups, although a black elite middle class is starting to establish its self. This paper attempts to explain why inequality persists in South Africa, and articulates suggestions for related research to gain a better understanding of the trajectory of poverty in the country.

During the past decade considerable efforts have been undertaken to address poverty and to overcome inequality in South Africa. This effort has been enhanced specifically with the dawn of democracy in 1994. The South African Government, with its Reconstruction and Development Programme (RDP) and Growth, Employment and Redistribution Strategy (GEAR) formulated programmes and projects within these macro economic policies to address poverty amongst the previously disadvantage groups and rural communities. Although significant progress has been made with regard to poverty reduction, inequality in South Africa still exists. The gap between the have and have-nots has become bigger as illustrated by the Gini-coefficient and income shares. After the International Conference on Population and Development in Cairo (ICPD) in 1994, South Africa changed its focus with regard to population and development. This paradigm shift also took place with the change of government. The focus was a movement away from reproductive health to a more human orientated approach, which addresses population concerns within the context of national development strategies. Poverty is one of these concerns.

This paper will attempt to illustrate efforts by the South African government to address poverty from 1994-2003 in line with the population policy that was adopted in 1998. It will also attempt to explain why inequalities still exist as well as present some recommendations to speed up the process with regard to poverty reduction in South Africa.

The population policy adopted in 1998 is designed to provide a comprehensive and multi-sectoral framework for addressing population issues that are currently considered not commensurate with achieving sustainable socio-economic and environmental development. A basic tenet of this policy approach is that the population concerns are considered as multi-faceted and intersectoral. It forms an integral component of national strategies for reducing past inequities, while substantially enhancing the quality of life of the entire population.

The policy emphasizes the shift to a sustainable human development paradigm, which places population at the centre of all development strategies, and regards population as the driving force and ultimate beneficiary of development. The role of population in development is encapsulated in the Programme of Action of the International Conference on Population and Development (ICPD) agreed upon by the international community in Cairo in 1994.

The goal of the policy is to bring about changes in the determinants of the country's population trends, so that these trends are consistent with the achievement of sustainable human development.

The policy outlines seventeen major population concerns that could constitute obstacles to sustainable development. The major population concerns concerning poverty include:

The policy objectives reflect the two main pillars on which the policy rests. One pillar is the systematic integration of population factors into all policies, plans, programmes and strategies aimed at enhancing the quality of life of the people at all levels and within all sectors and institutions of government. The other pillar is a co-ordinated, multi-sectoral, interdisciplinary and integrated approach in designing and implementing programmes and interventions that affect major national population concerns. Underpinning these two pillars is the need for reliable and up-to-date information on population and human development to inform policymaking and programme design, implementation, monitoring and evaluation. In essence focus has to be put on fighting poverty and inequality that have been brought about by apartheid

Up-to-date, national-level inter-temporal comparisons of poverty are rare at the moment. Among the products of Statistics South Africa, the census and the October Household Survey (OHS) are not suitable data sources, because they do not accurately establish household income.^[1] StatSA's Income and Expenditure Survey, which was conducted once in 1995 and again in 2000, is a vast improvement. However survey results for the 2000 round have not yet been released. Many recent analyses of the incidence of poverty still cite figures calculated on the basis of the 1995 Income and Expenditure Survey.

The most useful survey for our purposes would therefore appear to be the annual All Media and Products Surveys (AMPS) conducted under the auspices of the South Africa Advertising Research Foundation. Based on data from these surveys, Schlemmer (2002) assembled the following series of 'headcount' poverty measures, based on the household poverty line of R400 in 1989 Rand, and disaggregated according to population group.

The table reveals that the proportion of African and Coloured households living below the poverty line has increased substantially from 1993 to 2001. It also reveals that racial disparities in the incidence of poverty are still large.

Also based on the AMPS data, Schlemmer (2002) notes that there has been an expansion of the African middle class and a partial closure of the gap between average white and average African incomes. This leads to the conclusion that inter-race inequality is diminishing while overall inequality is as great or greater.

From a regional perspective, the most rigorous and illuminating example is the KwaZulu-Natal Income Dynamics Study (KIDS), in which about 1200 African households in KwaZulu-Natal that had been interviewed as part of the 1993 South African Labour Development Research Unit (SALDRU) survey were re-surveyed in 1998. This study reveals a similar trend in the incidence of household poverty. According to KIDS, between 1993 and 1998 the proportion of households in poverty increased from 33.7% to 41.6%. The headcount measures are quite different from those based on the AMPS data, largely because the poverty line used as a reference was a bit higher in the latter.^[2] What is significant, however, is that the direction and magnitude of change is quite similar, although not fully comparable because there is no indication what has happened in the three or four years since 1998.

The objective of this section is to present a picture of the diverse aspects of poverty in South Africa. There are three sub-sections. We first scan major poverty 'dimensions', by which is meant salient facts regarding the incidence and nature of poverty. Then, we discuss qualitative/experiential aspects of poverty, making the point that poverty is not simply a question of not having enough money. Finally, we identify categories of people who are especially vulnerable to becoming stuck in poverty.

Money-metric indicators of poverty are helpful in appreciating the incidence and distribution of poverty, but tell only part of the story. Over the course of the past two decades, there has been increasing attention to the fact that poor people's experience of poverty involves a great deal more than inadequate income or consumption (Chambers, 1994; Kanbur and Squire, 1999). This has been forcefully illustrated in South Africa by means of a number of qualitative research initiatives, of which perhaps the most ambitious is the South African Participatory Poverty Assessment, or "SA-PPA" (1998). We draw on the SA-PPA and other sources to describe some of the key experiential aspects of poverty, and where possible indicate what the trends are with respect to these.

Poor people are vulnerable to a number of harmful and potentially devastating threats, which they may not have the resources or power to avert. Among such threats are fire (e.g. shack fires, which destroy one's home and possessions); floods (because poorer people often end up erecting their shelters in flood-prone areas); job loss; crime (theft of money and possessions, and bodily harm); poor agricultural conditions (e.g. for those who rely in part on food production for sustenance); and illness and death in the family (often with no resources to seek medical care).

The experience of vulnerability is an aspect of poverty in and of itself, that is, a palpable disturbance to one's 'peace of mind'. In the SA-PPA, this sense of vulnerability was articulated in a number of different ways. For example, many participants expressed concern for their ability to cope with unpredictable crises. Bedford (1995, cited in May et al., 1997) notes the extreme stress and anxiety suffered by street children, who are exposed to constant threats of violence and sexual abuse. Chopra and Ross (1995, cited in May et al., 1997) indicate that African women in northern KwaZulu-Natal were particularly aware of the threat of children falling ill during certain times of the year.

South Africa's migrant labour system developed in distinctive ways owing to a combination of the singular importance of employment on mines, influx controls which limited the mobility of non-employed family members, and inadequate land resources left to black people on account of the inequitable land distribution. We focus here on one particular consequence of this system, namely a large number of effectively female-headed households.

In its 'ideal' form, the migrant labour system means that husbands earn an income in relatively lucrative jobs outside of the labour reserves rural areas, which they remit to their families back home and/or save. While this may often happen as planned, it frequently does not. First, migrant workers may lose their work, and thus have nothing to remit. Some of these industries, especially mining, are highly dangerous, and over the years many workers have sustained disabling injuries, succumbed to occupation-related illnesses,^[3] or been killed. Second, migrant workers may find other uses of their earnings, and may send remittances intermittently and insufficiently. From OHS 1999, we find that 21% of all households have one or more family members who are considered migrant workers, of whom 19% send remittances back to the household no more than once per year.

The SA-PPA relates the experience of many disillusioned women whose male partners failed to send remittances according to plan (May et al., 1997, p.37). Bank's study of a community in the Eastern Cape as part of the SA-PPA, notes that: "In DuncanVillage, we discovered that the classic formula for chronic impoverishment for female-headed households seemed to be the absence of a reliable male breadwinner in two consecutive generations" (1995, p.69; cited in May et al., 1997). Many of these women have the worst of both worlds. On the one hand, they receive inadequate or unreliable support, while on the other hand, by virtue of cultural norms, they often lack rights (e.g. land ownership), independence (e.g. to make own economic decisions), and voice (see below).

An important development over the past two decades has been the increasing frequency with which children have been left in the care of their grandparents, and in particular grandmothers, on rural homesteads. This reflects the increased participation of women in the urban formal and informal sectors, and the need to have someone look after their children while they do so. While this arrangement may benefit the household economically as a whole, it may also put undue pressure on elderly people, whose old age grants end up being used effectively to support whole households rather than just themselves.

Improved access to infrastructure, services, and amenities, is of course a major challenge of the post-apartheid government. The cholera epidemic currently rampaging through KwaZulu-Natal is a vivid example of the human costs of the infrastructure backlog.

A formidable amount of information is available on the extent of people's access, including disaggregations by race and place of residence (e.g. Stats-SA, 2000; May et al., 1995); on improvements since 1994 in this access (Khosa, 2000); on the problems with maintenance that have been discovered in respect of new infrastructure, e.g. water pumps (Department of Housing, 2000); on the capital costs associated with remedying the delivery backlogs (e.g. Jackson and Hlahla, 1999); and on the institutional mechanisms that must be put in place in order to accelerate delivery and ensure sustainability (e.g. Stavrou, 2000; Jackson and Hlahla, 1999). The number of different kinds of infrastructure and services is of course large, including electricity, water, communications, education, the judicial system, health care, sanitation, housing, refuse removal, financial services, and so on.

An important point, however, is that notwithstanding large strides in improving infrastructure and access to services and amenities, even when services are present, lack of access to them may still be experienced by the most marginalized members of communities. The SA-PPA shows how some community members’ ability to access services is hampered by a range of factors, some of which have to do with the design of the services themselves, and others of which relate to the extremity of people's poverty. For example, even where a health clinic does exist in the vicinity, and where it may offer free basic care, travelling costs may still be prohibitive, and people may find it difficult to be at the clinic during the times when it is open (May et al., 1997, p.61). Employees such as farm workers may hesitate to miss work in order to visit a clinic, for fear of being retrenched (Simbi and Aliber, 2000). Foreign workers who are undocumented migrants may be denied access to health services by xenophobic clinic staff (Mail & Guardian, 25 September 2000).

Crime and violence contribute to the experience of poverty at two levels. On one level, the exposure to crime and violence directly detracts from the quality of life of its victims and those fearful of being victimised. On another level, the high incidence of crime and violence that forms a salient feature of everyday life in South Africa, is symptomatic of a profound social malaise, wherein the cycle of poverty and violence are indistinguishable.

Crimes such as burglary and robbery can result in poor people losing what little assets they have (May et al., p.18), while the prevalence of violence adds to people's sense of vulnerability and oppression. While state-sponsored violence ended with apartheid and political violence has greatly subsided, violence among people who know one another in poor communities is rife, and is often linked to substance abuse. A 1996 study of homicide in the Eastern Cape found that 93% of all cases were linked to alcohol and drugs, while in Northern Cape, research had similar findings, and found as well that most cases were related to family disputes (CIAC, 1997, cited in Hamber and Lewis, 1997).

Violence and crime are increasing in rural areas (Hamber and Lewis, 1997), and in fact homicide rates are higher in South Africa's rural areas than in its urban centres (Hamber, 1999). That poor people are more likely to be victims of violent crime is borne out by the statistics, which show that a poor person is 80 times more likely to be injured or killed through violent crime than a wealthy person (Steinberg, 1999, cited in Hamber, 1999).

Women and children are especially likely to be victims of violence. Women face abuse by partners, while women and girls are subjected to a high risk of being raped. The incidence of rape is already high according to official police statistics (119 per 100 000; SAPS, 1999), but women's rights groups estimate that only 1 in 20 to 1 to 35 rapes are reported (People Opposing Women Abuse, 1995, cited in Hamber and Lewis, 1997). The highest incidence of rape is in rural areas (Hamber and Lewis, 1997).

Some studies suggest that the growing incidence of rape over the past 15 years or so, and particularly of gang rape, are aspects of a broader 'culture of violence', which in turn is a function of the protracted marginalisation experienced by many young men (Vogelman and Lewis, 1993). Arguably, the fact that economic opportunities accessible to many young men have not improved since 1994 despite the political watershed of that year, has served only to exacerbate the experience of marginalisation (ibid.). It is noteworthy that, of all South African men between the ages of 18 and 40, approximately 2-3% are presently in prison.

A study of perceptions of crime in Alexandra township in Gauteng (Stavrou, 1993), stressed residents' view that unemployed youth in search of an identity and status (e.g. through gang allegiance) are the most apt to commit violent and other crimes, and that this was aggravated by the rapid influx of residents in the township. Residents' quality of life was seriously affected by the prevalence of crime, not least because its perceived randomness means that it is difficult to take precautions to avoid it.

One other aspect of the experience of poverty as revealed in the SA-PPA, is that of lack of 'voice', which is a growing area of attention internationally. Broadly, a person 'has voice' when she feels she has an opportunity to somehow participate in decisions that may affect her life, as well as having avenues to lodge grievances with relevant authorities and institutions if she so wishes. Lack of voice was of course an acute aspect of many people's experience under apartheid, and was reversed in a significant way with the first all-inclusive democratic elections in 1994. The importance attached in the new dispensation to local government is a further reflection of the new government's earnestness to allow people these means of expression.^[4]

However, much of the experience of voicelessness in South Africa, particularly among the poor, is not at the formal political level, but at the level of communities and households. In their study of trends in public participation, Roefs and Liebenberg (2000) note that the majority of South Africans have little understanding of the role of local councils and parliament (80% and 73%, respectively), and only a minority of poor people specifically participate in any way in local councils or public hearings (23% and 22% respectively). However, 46% of poor respondents surveyed reported that they participate in some sort of community activity or organisation, and 30% of respondents indicate that community organisations are the most appropriate venue for addressing "problems in the community". While these figures are encouragingly high, they underline the fact that more marginal members of these same poor communities are apt to have an acute sense of social exclusion and voicelessness in their communities. This comes through vividly in the studies conducted as part of the SA-PPA. Chopra and Ross (1995, cited in May et al., 1997), for example, led their community participants in an exercise to identify the distinguishing characteristics of the poorest homes in their village in northern KwaZulu-Natal. Amongst the five characteristics identified, low income was not included, but not being part of community gardens or crèches was.

Social exclusion is experienced in distinct ways by different sub-groups. For the elderly, for example, the experience of social exclusion can be the result of fraught or absent family relationships. Maphorogo and Eager's study in Northern Province (1995, cited in May et al., 1997) identified that "a 'bad' or 'rude' daughter-in-law is a main cause of unhappiness of older people", as was living without one's spouse. Poor support generally from one's children or partner is thus considered emblematic of poverty. Apart from the fact that loneliness diminishes one's sense of well-being, the absence of supportive family members can attenuate one's links to the community, and render more difficult tasks such as collecting one's pension on pension day, etc. A recent report by the Ministerial Committee on Abuse, Neglect and Ill-Treatment of Older Persons (Department of Social Development, 2001), draws a frightening picture of the insensitive and/or exploitative treatment to which some elderly people are subjected. In addition to the abuse directed at some elderly people by their own family members, the report highlights poor conditions in residential homes, at pension payout points, and at clinics.

People with disabilities face forms of social exclusion, which contribute directly to their likelihood of missing out on economic opportunities. This social exclusion operates independently of the actual physical or mental challenges that the physically challenged may face. As such, 'disabilities' are in large measure a social construct, which has the effect of excluding certain individuals (Yeo, 2000).

One can fairly ask what personal attributes or environmental factors tend to keep people in poverty. The circumstances in which one finds poverty are in fact quite diverse. In addition to unemployability and poor business prospects, poverty can be a function of discrimination, or due to the severance of social networks, or in still other cases it can be the outcome of complex processes such as the migrant labour system. This subsection describes some of the main categories of poor people in South Africa, and attempts to relate this poverty to underlying causes and influences.

Poverty is especially prevalent in rural areas, and particularly among Africans and Coloureds. The national data on poverty and unemployment, together with some of the statistics derived from the KIDS data, indicate that, as a group, rural blacks have a high chance of being poor. Lack of access to employment is arguably the single greatest cause of rural poverty. In 1999, over 51% of the rural African workforce was unemployed, versus 43% for Africans in urban areas. However, rural poverty is aggravated by lack of access to productive resources. According to the Rural Survey (Stats SA, 1999), in 1997 there were some 900 000 African households living in former homelands that had no access to arable land, 1.4 million that had no livestock other than chickens, and 770 000 households that had neither.^[5] This is largely a function of the overcrowded conditions in these areas. For those people dwelling in former homeland areas that do have access to arable and grazing land, that access can make a significant difference, though it does not often offer a route out of poverty. Based on the 1993 SALDRU data, it was estimated that among the poorest 20% of households in former homeland areas having access to arable land, agricultural production contributed 35% of total income (LAPC, 1996). Moreover, as Ardington has shown (1988), poor households in particular rely on selling livestock to offset financial crises. Shackleton, Shackleton, and Cousins (1999) have shown that beyond conventional land-based products such as grains, vegetables, meat and milk are numerous other valuable products that can be harvested or derived from the land, whether their own land or from the commons.

From the 1999 October Household Survey, 42% of all African households, i.e. 2.7 million, are female-headed. By and large, these can be considered single-parent households, though they may occasionally receive remittances from absent males.^[6] In one third of these households, the household head is the only adult (18 years and older) in the household. Around 17% are so-called 'granny households', i.e. the female household head is the grandmother rather than the mother of the children in her care. The reasons why female-headed households have a high probability of being stuck in poverty are numerous. The most obvious reason is that many female-headed households rely only on the income of the mother, or worse, she may not have an income at all. She may rely mainly on child support grants from government, or on remittances from relatives or gifts from benefactors. As likely, the household head does have some form of employment, but there is a good chance that it will be self-employment or employment in the secondary labour market with a low level of remuneration and under poor conditions.

According to the 1996 census, 2.7 million people in South Africa have disabilities (Stats SA, 2000).^[7] Of these 2.7 million people, 1.6 million are adults between the ages of 20 and 65. Disabilities can reduce one's chances of obtaining a job, can impede one's pursuit of self-employment in the informal sector, and can also impose medical and other costs that one would not otherwise have to bear. Of course, being physically challenged does not imply that there are no other household members who earn decent incomes. Surprisingly, data from the 1996 census show that there is not a strong link between disabilities and unemployment. The unemployment rate for Africans with disabilities, for example, is 45.3%, which is only marginally above the unemployment rate (broad definition) for Africans for that year of 42.5%. Likewise, for Coloured people with disabilities, the unemployment rate is 26.7%, versus 20.9% for all Coloured people. Moreover, the occupational profiles of the physically challenged are not very different from those for the general population. Presumably in a tighter labour market the discrepancies would be more stark.

Elderly people are a specific concern in terms of poverty especially if, like the physically challenged, they must fend entirely for themselves and for dependants. Many elderly people who do not have others on whom to rely on for support do receive old-age grants, but it is clear from Roberts (2001) that these grants are in no way sufficient to keep a household out of poverty. In other words, the widespread allocation of old-age grants is not an indication that there are no poor households effectively headed by elderly people.

Agricultural employment peaked around 1968-1970 at over 1.6 million workers, of whom 99% were African, Coloured, or Indian (Department of Agriculture, 2000). It is not possible to say how many of these were regular, and how many were seasonal or casual workers. Very likely around 8%-15% were casual or seasonal, and the rest regular. In 2000, the total number of regular farm workers was around 580 000, implying a decline over the past three decades of some 860 000 jobs, of which the majority were almost certainly through retrenchments. Bekker et al. (1992) argue that former farm workers are particularly susceptible to social exclusion. The reason is that, having been retrenched, farm workers are usually evicted from the farms where they have resided. Many retrenched farm workers were in fact born and raised on the farm from which they are eventually ejected, thus they have no 'roots' elsewhere to which to return. (In principle these types of evictions should have ended with the Extension of Security of Tenure Act Bill of 1997, but in practice this has often not been the case.) Retrenched farm workers are thus severed from their existing social network – e.g. other farm workers in the immediate area – and forced to settle, typically with little or no savings or other capital, in townships, squatter settlements, or communal areas. Some attempt to switch to seasonal agricultural employment, as other employment prospects may be even poorer, given the non-transferability of farm workers' skills.

Refugees, asylum-seekers, and undocumented migrants face particular kinds of exclusion and deprivation in South Africa. The vast majority of these people come to South Africa from neighbouring African states in search of economic opportunity. Some migrants indeed manage to improve their circumstances significantly compared to their country of origin, particularly those with entrepreneurial skills (CASE, 1998). The number of undocumented migrants is by definition difficult to gauge. Estimates range from 250,000 to eight million (National Population Unit, 2000; CASE, 1998). If the trend observed among visa-overstayers is any indication, of whom there were an estimated 800 000 in 2000, then 70% to 80% have probably arrived in South Africa since 1994.

'The homeless' is an amorphous category. This is particularly so in South Africa due to the fact that there exist hundreds of thousands of people living in informal squatter settlements who do have homes, but whose homes are obviously very unsatisfactory. Many are former farm workers, as discussed above. Many are wage earners in urban areas who cannot afford decent housing there, and for whom the costs of commuting into and out of the city are a significant burden. Their solution may be to pay rent in overcrowded flats in urban slums (Central Johannesburg Partnership for the Homeless, 2001). Those homeless living 'on the street' would appear to be a somewhat different category, though there may not be a clear line separating the two. These people, who are often referred to as the 'street homeless', are typically lone individuals or children rather than family units, and have severed or lost ties with social networks they may once have had.

AIDS orphans are defined by UNAIDS as children under the age of 15 who have lost their mother or both parents to AIDS. The number of AIDS orphans is set to rise as South Africa's high HIV prevalence rate among adults translates into a higher prevalence of AIDS and then AIDS deaths. UNAIDS estimates that at the end of 1999 there were around 371,000 living AIDS orphans in South Africa (UNAIDS, 2000,), while 50,000 AIDS orphans have already died (presumably from AIDS but also other causes, as HIV-negative AIDS orphans have a higher mortality rate than non-orphans). The Metropolitan Life model estimates that by 2005 there will be 920,000 AIDS orphans in South Africa, and by 2010 there will be roughly two million (reported in Whiteside and Sunter, 2000). By contrast, according to the 1996 census, the total number of motherless orphans 14 years and younger in the country was about 400,000. This figure is presumably inclusive of AIDS orphans of that time, but at any rate the number of AIDS orphans will soon account for a very large increase in the total number of orphans in the country.

Left untreated, adults who become infected with HIV develop symptoms of AIDS within 6 to 8 years, and most die within 10 years. Treating HIV to delay the onset of AIDS and opportunistic infections means a longer life, but in the absence of government support this will generally be affordable only to those who are relatively wealthy. For everyone else, the economic effects of the infection will mainly occur when one develops AIDS, from which point the economic effects on the AIDS sufferer and his or her family can be devastating. A study of urban households in Côte d'Ivoire, for example, showed that the average decline in household income when a household member becomes ill with AIDS is 52% to 67%. Meanwhile, household health expenditure quadruples and food consumption declines by 41% (UNAIDS, 1999). A study by the HSRC on AIDS-affected households in rural KwaZulu-Natal supports these findings.

The extent of inequality in South Africa is well known. Figure 1 shows the distribution of the population across percentiles of the expenditure distribution. The black population completely dominates the lower percentiles of this distribution, while the white population is concentrated in the upper percentiles, particularly above the eightieth percentile.

But even in 1995 – before the full consequences of political change had been felt –, a substantial proportion of the black population already found themselves in the top two deciles of the expenditure distribution. To summarise 80% of the black population with regard to expenditure are still falling within the lower ranks of expenditure while the white population still hold the upper ranks of the distribution.

Figure 2 shows a very large gap in per capita household expenditure between the different population groups in 1995. Yet interestingly, within the largely poor black population, the top decile had moved far ahead of their counterparts is illustrated by figure 3. Its per capita expenditure was even well ahead of that of the full Indian population, and starting to approach that of the white population. This trend is likely to have continued since the political transition, with the consequence that the top segment of the black population – the black elite – has almost caught up with the white population as a whole. This is also reflected in purchasing patterns in the market.

Thus inequality in distribution is mainly a problem of inequality between white and black, but inequality is no longer only a racial issue: the top part of the black population already exhibits relatively high levels of per capita household expenditure, although the largest part of the black population remains poor.

Given large-scale inequality in the market, the aspirations and expectations of the newly democratised society have focused on the role of government in the redistribution of income. In this regard, the budget is crucial. Figure 4 shows the percentage of social spending going to the different racial groups. Until 1993, apartheid bookkeeping made it relatively easy to obtain such spending figures, because social expenditure was clearly demarcated by race. Since then, however, this became less easy, because of the deracialisation of expenditure.

For this reason the Department of Finance launched an investigation into fiscal expenditure incidence since the transition, based mainly on fiscal data from the various social expenditure departments and household surveys. (Van der Berg 2000; Van der Berg 2001a & 2001b) This allows some analysis of trends in social spending by race since the political transition.

The figure clearly shows a considerable shift in fiscal incidence from the mid eighties onwards. This is not a historical anomaly – governments often respond to social pressures by first instituting hesitant social reforms, in an attempt to deflect pressure, before political reform comes onto the agenda. Thus the new government had relatively limited scope to shift fiscal incidence much further towards the black population, and particularly to the poorer part of the black population.

Figure 5 shows a Lorenz curve of income excluding social transfers. It shows the cumulative proportion of pre-transfer income from the poorest to the richest quintiles as well as concentration curves for the various types of social expenditure. Concentration curves show the cumulative proportion of spending going to these same groups:

Thus already by 1997, most social programmes were progressive in their effect (i.e. the poor gained more than proportionately from such spending relative to their pre-transfer income) and most were also relatively well targeted, i.e. the poor gained more than proportionately to their population size. Clearly, social spending in South Africa is already relatively well targeted towards the poor, or not as skew as in many other developing countries. This is particularly the case for social transfers, an important item of expenditure in the South African context, which are extremely well targeted. Tertiary education is least well targeted, mainly because of issues of access: few poor children fare well enough at school to enter tertiary education, and those who do are constrained by the high cost. In contrast, the fiscal benefits of school education were in 1997 already almost equally distributed across the income range. Despite remnants of past discrimination in per pupil expenditure, the poor have more children, and almost all South African children are now at school until about age 15.

The Lorenz curve is used to illustrate economic concentration in any particular population. It works from the assumption that a straight line represents a distribution where the poorest 20% also receive 20% of the income or economic advantages in a society (or 50% of population receives 50% of the income). Any departure from the straight line indicates economic concentration of wealth in the hands of a smaller proportion of the total population.

Figure 6 shows a Lorenz curve for income from which has been subtracted all transfer incomes, in order to investigate the impact of government spending and taxes. The Gini coefficient associated with this Lorenz curve is 0.66, a very high level indeed. If taxes are also included, the Gini coefficient declines only slightly to 0.64 per cent. If the impact of social spending is also included, however, in other words a wider concept of income is employed, the Gini declines to 0.44. This illustrates very clearly that the redistributive impact of taxes is far less than that of social spending, something quite common in developing countries, where taxes play a small role in overall income distribution.

The column second from the right in Figure 7 shows the geographic distribution of the population between metropolitan areas, other urban areas, and rural areas. The distribution of total social spending (the final column on the right) closely follows this geographic distribution of the population. This is exceptional for a developing country. Developing countries usually experience considerable urban bias in public spending, sometimes for reasons related to strong urban pressure groups, but also because of difficulties of rural social service provision and lower rural participation in education. But in South Africa close to universal access to education, fairly good targeting of social old age pensions and disability grants, and opting out of public health services by the affluent have contributed to surprisingly good targeting of rural areas.

Individual social spending items do show some stark deviations from the overall pattern. So, for instance, housing spending is disproportionately targeted at metropolitan and even other urban areas, and does not reach rural areas. In contrast, social assistance spending (social transfers) is extremely well targeted at the rural areas, which get 56% of such spending, as against their 42% population share.

Thus South African fiscal incidence is quite favourable to the poor and to rural inhabitants, as far as most aspects of social spending are concerned. This partly results from the fact that a large part of the most affluent population, particularly whites, do not partake in many social spending items to the same extent as their share in the population. For instance, not many of the rich use public health services, particularly clinics, nor do many get social transfers. Moreover, the affluent usually have fewer children who can benefit from public school spending.

This section has two goals. The first is to document the incidence of unemployment and to describe changes in the quality of employment over time. The second is to establish a link between employment status and one's quality of life. In both cases, statistics are derived mainly from the October Household Survey (OHS) of 1999.

Unemployment is high and rising in South Africa. Between 1994 and 1998, the aggregate unemployment rate rose from 20% to 26%, according to the "narrow definition" used by Stats SA, whereby a person is unemployed if he or she has actively sought work in the past four weeks. According to the "broad definition" of unemployment, whereby one wishes to be employed but has not necessarily sought work in recent weeks, the change between 1994 and 1998 was from 31% to 39%. As Kingdon and Knight (2000a) point out, the gap between the narrow and broad definition figures suggests that a large fraction of the work force is 'discouraged', that is, have ceased actively searching for employment out of frustration, or for lack of resources, or both.

Table 2 shows that the incidence of unemployment is considerably greater for Africans and Coloureds than for Whites. It also disaggregates employment by 'sector', by which is meant here informal and formal, and employment versus self-employment. Overall, Africans and Coloureds are more likely to be employed or self-employed in the informal sector, and in absolute terms whites comprise a relatively tiny share of activity in the informal sector.

Not surprisingly, the incidence of unemployment is strongly differentiated according to race, gender, and place of residence. In 1999, the incidence of employment and unemployment was as follows, where unemployment is understood according to the broad definition^[8]:

Unemployment is clearly more severe among women than among men, and among Africans it is considerably more severe in rural areas than in urban areas as reflected in Table 3. The fact that unemployment among Africans is higher in rural than in urban areas, is something of an anomaly, and probably in large measure is due to the insufficiency of African's access to agricultural land as reflected in Table 4. this is also reflect that the majority of African and coloured populations fall in the poor and ultra poor category. Part-time and casual employment make up relatively small shares of employment, though it must be noted that the survey picks up those who have been casually employed within the previous week, as opposed to those who may have had casual work within, say, the year.

In addition to the static picture presented above, we note the following salient facts (following Kingdon and Knight, 2000b), based again upon data from the 1999 October Household Survey:

The situation is complicated by the changing nature of formal sector employment in South Africa. The overall pattern of formal sector employment in South Africa over the past several years is that fewer people have employment but those who do have enjoyed real increases in remuneration. However, there is also a broadening grey area involving more employment through the secondary labour market, whereby tasks that had previously been performed by permanent, regular workers are increasingly being performed by temporary, casual, and part-time workers, or what are sometimes called "independent contractors". The decline and 'flexibilisation' of employment in the commercial agricultural sector are especially damaging for rural livelihoods, because commercial agriculture represented almost one third of formal sector employment in rural areas in 1998, and yet the total amount of regular employment on farms has declined by about 20% over the past decade.

The link between employment type and unemployment on the one hand, and personal or household welfare on the other, must be established. One way of doing this would be to calculate headcount poverty measures for different categories of employed and unemployed households. Unfortunately, as noted above, the OHS is not suitable for this task, while the author does not have the AMPS data on hand. We therefore use a proxy for household welfare from OHS 1999, namely whether or not a household reports not having enough money to purchase sufficient food for its children at some point during the year.^[9] The virtue of this measure as a proxy is that it focuses on a feature of welfare that is universally considered important – the ability to eat. The downside is that this particular measure applies only to those households with children aged 7 or younger.

Figure 8 below shows the frequency with which respondent households indicated that they had not had enough money for food, according to the employment status of the household head or acting household head.

One problem with the picture presented above is that it does not take into account the fact that household members other than the head or acting head may also perform an important role in supporting the household. If we look at the employment status profile of households (excluding households in OHS 1999 that comprise only one person), we find that 42% of all African and Coloured households have no member with formal sector employment or self-employment; only 19% of those without a member in formal sector employment or self-employment have a member engaged in some form of informal sector employment or self-employment; and 34% have no member with any form of employment or self-employment at all.^[10] One would be mistaken in thinking that households lacking employed members tend to be small relative to those having employed members – in fact, a disproportionate number of larger households are exactly those households in which there is no member engaged in the formal sector or in which all the adults are unemployed. Presumably these are households that rely on a combination of social grants and remittances, e.g. 'granny households'. The implication of all this is that unemployment and under-employment tend to run in households, and one would expect these to be the most vulnerable households of all.

This section starts with a macro-view of anti-poverty initiatives and the resource flows associated with them. The following three sub-sections then describe three distinct types of anti-poverty initiative, partly because these are important in themselves, and partly because they are emblematic of problems besetting the fight against poverty generally.

The following matrix summarises the main types of poverty alleviation initiatives and programmes (Table 5). The purpose of the matrix is to indicate the typical activities of government, private sector, and other types of civil society entities, in terms of the broad categories of social security and promoting development. The table also ventures some rough order-of-magnitude estimates as to the average annual expenditure associated with each sector for the main types of initiatives. (To avoid double counting, for NGOs and CBOs, these estimated expenditure figures do not include what originates from the government or private sectors.) Although expenditure is not synonymous with impact, the figures to convey a sense of which sectors are and are not relatively significant in these two broadly defined areas of intervention.

Three conclusions can be drawn. First, spending on social security is many times greater than spending on development and job creation. Second, spending on social security to low-income households comes predominantly from government. And third, although spending on development and job creation is relatively modest, that also is largely dominated by the government sector.

One implication of these observations is that a major reorientation of spending may be necessary if government is to truly foster 'developmental welfare'. The obvious question is, how can this be afforded given the absolute necessity of leaving the safety net in place? A more searching question, of course, is whether this vision has merit in the first place. In the sub-sections that follow, we make the case that, at any rate, the pattern of expenditure revealed above does not appear to correspond to a credible short or medium-term strategy for addressing South Africa's massive inequalities.

There are three main public works programmes operating in South Africa: the Community-Based Public Works Programme, Working for Water, and LandCare. The programmes have accomplished a great deal and in many instances have generally been managed well.

The programmes have accomplished a great deal and in many instances have generally been managed well. Expenditure on public works programme has increased almost tenfold since 1998. As part of the objective of providing employment, the public works programme has employed a total of 124 808 people since 1998. However, most of these jobs were temporary in nature. A total of 3 407 permanent jobs were created in 1999-2002 and a total of 141 permanent jobs were created by the first half of the 2002/03 financial year. Research indicates that public works programmes vary in heir efficiency of transferring income to the poor with the average expenditure per worker varying between R 27 242 in Limpopo to R 6 515 in the Eastern Cape.

Land restitution, tenure reform and land redistribution are the other major policy instruments used to alleviating asset poverty since 1994. Since 1994, 1.8 million hectares (ha) of land have been transferred under the redistribution programme to about 137 478 households. Approximately 80% of these transfers occurred between 1997 and 2002. By 2002, approximately 68 000 claims had been lodged of which 72% were urban and 28% rural. A total of 36 489 claims have been settled involving about 85 000 households.

South Africa’s experiences and lessons learned in implementing population, environment and development policies has demonstrated that community based initiatives, which resonate with people’s basic needs, can make a huge difference among poor communities. A community-based environmental and reproductive health programme in two rural districts was initiated in 1998 by the government, together with UNFPA, the Planned Parenthood Association of South Africa and the Working for Water Programme. What is notable about the programme is the overt linkage of population interventions to an environment and development programme with beneficial effects to the communities through the provision of clean water, job creation and promotion of reproductive health information and services, including HIV/AIDS. The project, which was undertaken to restore original water flows to rivers and streams, created many jobs, especially for women, and then became linked to the provision of project based reproductive health services.

At the community level, the effects have been both positive and negative, as is evident in a recent review of secondary information on Working for Water and LandCare (Aliber, 2002). First and foremost, the participation of a large fraction of a community's members in these programmes as wage earners means a large infusion of cash into the chosen community. This has allowed participants to pay off debts; eat more and more nutritious food; repair dwellings and replace worn-out clothes; invest in livestock and fencing, and pay for tractor services; stock small businesses such as spaza shops and shebeens; and afford long-delayed medical attention. Spillovers to non-participants have included boosted sales of vegetables to participants who no longer have time for gardening; increased firewood sales, including to adjacent communities; and less danger of stock theft due to fewer trees impeding supervision of one's animals.

A less auspicious effect ascribed to the programmes has been the failure of numerous income-generating activities that had been at a formative stage when the programmes were introduced. The failure of these projects is apparently due to the fact that those who can get employment on a project do so, even if it means abandoning another activity, simply because the daily wage through the project is so high by local standards. Also, rather than investing more in agriculture, many households apparently stop ploughing altogether, and others build up new layers of debt, not least through hire-purchase contracts with furniture shops.

On a macro level, however, one draws a different conclusion as to the poverty reducing effects of the programmes. At the conclusion of financial year 2000/01, roughly 50 000 people were employed under the three public works programmes in that year. To put this number in perspective, the total number of 'jobs' created under the three programmes represents about 1.7% of the estimated 2.9 million unemployed rural Africans and Coloureds. Bearing in mind that the number of regular farm worker jobs that have been lost in an average year since 1990 is 14 000, the public works programmes collectively are insufficient in scale to stem the tide of increasing rural unemployment, much less reverse it. Although the employment generating aspect of the programmes is certainly laudable, the programmes cannot be said to be having a palpable effect on poverty reduction in a direct sense, except for the relatively few individuals or communities who have had the good fortune to be selected.

There are indications that government is considering expanding public works programmes, which, notwithstanding some of the misgivings indicated above, is probably a worthy way of contributing to livelihoods in poor communities. The figures above suggest however that the programmes will have to be scaled up by a factor of 10 at least to make discernible inroads into poverty, and will have to be sustained if positive spin-offs will be generated and translated into sustainable growth.

StatsSA's reported employment trends based on consecutive OHS 1996 and OHS 1998 purported to show that employment levels had stabilised, but mainly because formal sector job losses were roughly off-set by increases in informal sector employment and self-employment. While the fact that these consecutive surveys do not comprise a panel and do not allow us to know for certain, it seems likely that many of those entering self-employment in 1998 were those who lost formal sector employment in 1996 and 1997.

Notwithstanding the fact that informal sector employment is inferior to formal sector employment, from both an individual and societal point of view it is preferable to unemployment. Kingdon and Knight (2000b) marvel at the fact that the share of informal sector employment in South Africa's economically active population is so low, despite such high levels of unemployment. Drawing again on Table 2, we see that among Africans and Coloureds in 1999, there were 6.7 million unemployed people compared to one million self-employed in the informal sector. By contrast, 90% of the India workforce are self-employed in the informal sector.

Kingdon and Knight therefore ask the question why the unemployed do not enter informal sector self-employment in larger numbers. They conclude that, in so far as the unemployed are much worse off than the informally employed across a range of indicators, then there must be barriers to entry. Among these barriers to entry are possibly lack of capital, lack of entrepreneurial skills, and inability to penetrate informal community networks that may control opportunities, e.g. by controlling access to street space or taxi routes. The fact that formal sector self-employment is on average more likely to ease household poverty than informal sector employment (Figure 1 is on distribution of individuals by race and expenditure decile), also suggests that there are barriers to entry that must be contended with. Unfortunately, quite possibly the biggest limitation on the SMME at this point in time is not an entry barrier but lack of an effective market.

Government is not well placed to foster the SMME sector. Thus far, it has attempted to do so in two ways, indirectly through the creation of institutions that render support to the SMME sector (e.g. Khula and Ntsika), and directly through the promotion of income-generation projects. Neither strategy has proven very effective, and the direct promotion of income-generation projects has been downright discouraging. The reasons very likely have less to do with whether or not government is 'doing its job well', than with the question whether this is the sort of function that government is suited to performing. As suggested above, there is reason to suspect that, no matter how assiduously government pursues its multi-pronged SMME promotion strategy, it is doomed to bear little fruit in many areas so long as disposable income in poor communities remains so low. As Ardington (1988) shows, for example, income-earning opportunities from self-employment tend to diminish when formal sector ("core economy") employment falls, since lower remittances to rural areas mean less disposable income fuelling the informal economy there. The decline in formal sector employment may have intensified the need to see the SMME sector thrive, but it has also imperilled the prospects for exactly that to happen. It is likely that the most efficacious SMME promotion strategy would be to ensure that disposable incomes are increased on a broad basis, whether this is through a massively scaled-up public works programme or the introduction of a basic income grant. But apart from this, it must be acknowledged that a more vibrant SMME sector will do little to remedy the country's massive inequalities unless it becomes much larger and more vibrant.

Land redistribution was initiated by the Department of Land Affairs in 1995 with a flat household grant that was gauged to match the new housing grant. The programme had multiple objectives, arguably the most central of which was to provide tenure security to rural households that lacked it. The economic case for the land redistribution programme was never made clearly, except in early World Bank inputs which argued that the move to smaller farms would result in a rise of aggregate production, a claim that could generously be described as fanciful.^[12] The arbitrary target of redistributing 30% of agricultural land within 5 years compounded the confusion, in the sense that it further distracted policy makers from thinking through how many people land redistribution should seek to benefit, and in what way.

Problems with redistribution projects launched in the first five years led to a major revision of the programme, resulting in the Land Reform for Agricultural Development (LRAD) programme, which was launched in August 2001. Many of the changes made to the programme made sense in terms of the analysis of what plagued the earlier generation of projects. However, there has been a failure to revisit the aims of redistribution and to realistically articulate in what way redistribution can benefit the rural economy or the rural poor. The meaningless target of redistributing 30% of the agricultural land was resuscitated, with little thought to what this would mean for rural employment or the rural economy.

LRAD is as yet quite new and it will take some time before it can be properly evaluated. However, a few comments can be ventured based on the types of projects that are emerging. Thus far, there appear to be two main types of projects. One kind of project involves members of the local elite accessing the grant in conjunction with loans from the Land Bank, in order to purchase a going concern. These projects involves a small number of beneficiaries, but a large amount of money from the government (on average R100 000 to R200 000 per household, against R16 000 under the old programme). The other type of project involves farm workers applying as a group to purchase the farm on which they have been employed. Projects in this second category involve a larger number of beneficiaries, often accessing smaller government grants per beneficiary. A large fraction of these projects are seller-driven, and the motives of the would-be seller are mixed. On the one hand, farmers exit farming all the time, and if one's children do not wish to take over then one is forced to sell. However, increasingly the decision to sell is motivated by concerns for personal security. In any event, owners perceive land redistribution as the best opportunity to sell their land, and the introduction of LRAD, with its much larger grants, has greatly facilitated their getting out of commercial agriculture.

Although in some cases the acquisition of an ownership stake will indeed benefit erstwhile farm workers, it is doubtful that the overall effects of the land redistribution programme can be said to make a significant positive difference to the rural economy. In the first place, there is little or no evidence of an intensification of land use on properties that have been transferred; and second, related to this, there is no net creation of new employment opportunities. To the extent LRAD will facilitate the emergence of a cadre of middle class black farmers, it is unlikely they will ever comprise more than a fraction of 1% of the total number of rural unemployed. Redistribution may have laudable motives, notably in redressing past inequities, but the inattention to developing an economic logic for redistribution represents a significant weakness. Ironically, land redistribution is attracting ever more resources (though still very modest as a share of the total government budget), at the same time that there is a growing problem of under-utilisation of land in former homeland areas.

This paper has attempted to make the case that the trajectory of poverty that obtains in South Africa is intimately linked to developments in the labour market. This is by no means a novel insight, but neither does it always receive adequate focus. The second message of this paper is that reversing the trend towards greater impoverishment will require a reversal of the decline of formal sector employment, and that other strategies (redistribution of land, promoting the SMME sector, etc.) will generally only help at the margins unless thoroughly modified. This relates to the third, more understated message, which is that government initiatives that in themselves are worthwhile often have little significance overall because the scale at which they are implemented is no where near commensurate with the scale of the problem they are trying to address. (Public works programmes are a case in point.)

All the above must be read in the context of the population policy that was adopted in 1998. The South African Population Policy that was adopted by the 1994 democratic government has shifted from family planning methods. It places population within the development paradigm, which was advocated by the ICPD in 1994. The focus is to fully integrate population concerns into all development strategies, planning, decision making and resource allocation, with the goal of meeting the needs and improving the quality of life of the present and future generations. Focus is especially placed on the status of women, specifically African and rural women, adolescent reproductive health and questions around poverty. The aim is to address these problems in an integrated manner.

The major thrust of the policy therefore is to ensure that births, deaths and migration factors are taken into account with all planning. As such, a survey for policy implementation was conducted during 2000. The results of the survey shows that most government departments supported the strategies proposed in the policy and that there is a shift in policy planning. A programme to strengthen population development system in government has been developed. This programme is a joint effort between Social Development, Statistics South Africa, (Government Communications Information System (GCIS) and United Nations Population Fund (UNFPA)

South Africa enjoys the advantage that there is a prodigious amount of policy-relevant research being conducted on poverty and poverty-related issues. Our understanding of poverty is improving constantly, and benefits enormously from the increasing store of large-scale surveys, such as those conducted by Stats SA and the SA Advertising Research Foundation. What follows are specific suggestions for research that relate to the themes touched on in this paper:

^[1]The OHS, which was issued each year from 1995 through 1999, did collect information on household expenditure; however this was done by asking a single question requiring the respondent to indicate in which band total monthly household expenditure fell. By contrast, best practice is to ask the household about its expenditure on a whole long list of items so that nothing is omitted.

^[2] A poverty line is typically based on some measure of what a household needs in order to ensure adequatefood and to enjoy a certain standard of living. An example is the 'household subsistence level' poverty line established by the Institute for Planning Research of the University of Port Elizabeth. However, any such poverty line is essentially arbitrary and must be understood as such. The well-known 'dollar a day' measure of poverty used by the World Bank is arguably even more arbitrary.

^[3] South Africa's Chief Inspector of Mines recently stated to the media that some 196 000 former mineworkers suffer from occupation-related diseases (SAPA, 19 September 2000).

^[4]In the words of the RDP policy framework document of 1994, "Local authority administrations should be structured in such a way as to ensure maximum participation of civil society and communities in decision-making and developmental initiatives of local authorities."

^[5] These figures represent 38%, 59%, and 33%, respectively, of the total number of households residing in the former homelands as of 1997.

^[6] The October Household Survey does not operate according to a formal definition of "household head", rather enumerators are instructed to interview the actual or "acting" household head, who is identified by members of the household themselves. The "acting" household head would typically be a woman whose husband is considered to be the household head despite being absent much or most of the time. A weakness of this approach is that it assumes a single head where the reality may be more complex, and makes no distinctions in terms of household decision-making, responsibility for economic sustenance, etc. (Rogers, 1980).

^[7] As Stats SA points out, the term "disability" can mean different things to different people as well as to different governmental statistics agencies, as clearly shown by the fact that official statistics for Australia indicate that 23% of all Australians have disabilities. For the purposes of the 1996 census respondents were asked whether any household member had a "serious sight, hearing, physical, or mental disability". One problem noted is that the census did not cover residents of institutions, including those that exist specifically to serve people with disabilities (Stats SA, 2000a, p.73).

^[8]It is important to note that these figures are inclusive of informal sector employment, which is readily captured in the October Household Survey (OHS). Also note that in general figures for 'Indians/Asians' are not reported in this section, as they comprise a very small fraction of the workforce participants captured in the Survey.

^[9] The exact question from OHS 1999 reads, "In the past year, was there ever a time when children under 7 years of age went hungry because there was not enough money to buy food?"

^[10] The balance of about 6% would appear to be households where no adult member is in the workforce.

^[11] Note that the figures may or may not be inclusive of personnel costs, depending upon whether the benefit consists primarily of cash transfers or of services rendered.

^[12] The claim is based on the belief by some academics that there exists an inverse relationship between farm size and productivity. While this is no doubt true in some contexts, that it would obtain in generally arid South Africa is implausible.

The following images related to this document are available:

Photo images