The situation in South Africa presents unique challenges to achieving sustained poverty\nreduction. Although it is an upper-middle-income country with a per capita income\nsimilar to that of Botswana, Brazil or Malaysia, a significant proportion of South African\nhouseholds have remained poor despite a plethora of government policies that target the\nless resourced. While estimates vary, over 22.9 million South Africans are categorised as\nbeing poor, with almost 2.5 million people suffering from malnutrition. Most analysts\nnow agree that while poverty increased during the 1990s, some progress has been made\nin reducing both the incidence and depth of poverty after 2000.\nThis thesis argues that the economic and social dynamics set in motion by apartheid that\nproduced this situation, may also have generated a low-level equilibrium trap from which\nsome the poor in South Africa will find it difficult to escape. The thesis suggests that the\nexplanation for this 'poverty trap' lies in what Sen has termed the exchange entitlement\nmapping that poor households face when attempting to use their assets/endowments. In\nother words, the processes that underpin the accumulation of assets, the opportunities to\nuse these assets, and the returns obtained are structurally prejudiced against the poor.\nThe implication is that the current experience of poverty leads to its reproduction and to a\nstructurally persistent poverty.\nThe central research question of this thesis is then: \"Did the extent, distribution and\nexperience of poverty of the apartheid era persist in the immediate post-apartheid South\nAfrica despite the efforts of government to foster pro-poor reforms?\" The central policy\nconcern is that if asset accumulation failure underpins persistent poverty, policies for\nthose who are structurally poor should be differentiated from that which is directed at\nthose who are transitorily poor. As an example, the policies of the South African\ngovernment concerning the redistribution of agrarian assets (principally land and finance)\nmay not be sufficient to assist the poor in rural areas, and may only increase intra-rural\ninequality.\nThe thesis draws on two principal data sources: the South African Participatory Poverty\nAssessment completed in 1997, and the KwaZulu-Natal Income Dynamics Study (KIDS)\nwhich contains panel data collected from the same households in 1993, 1998 and 2004.\nUsing these data, the thesis identifies a typology of structural poverty classes. At the\nbottom of this typology are those trapped in poverty with an asset base that is inadequate\nto meet their immediate needs as well as their ability to accumulate further assets over\ntime. Other are stochastically poor or non-poor, moving in and out of poverty according\nto their good or bad fortune. Finally some have never been poor and have the asset base\nto ensure that they remain in this position or indeed improve over time.\nThe livelihood strategies of households are used to differentiate households according to\ntheir participation in labour markets, farm and non-farm own production and access to\nsocial grants. The livelihood clusters that result are then matched to the poverty classes\nin order show differentiation among the households surveyed in KIDS. This allows for\nmore nuanced policy recommendations that can be tailored to the needs of households\nexperiencing different forms of poverty.