Philosophiae Doctor - PhD / State social security transfers for families existed in South Africa only in the form of state maintenance grants, which paid up to R700 to single parents. The system was not appropriate in the South African context, being racially biased and financially unsustainable. The Department of Welfare - following in principle the recommendations of the "Lund report" - introduced with effect from 1 April 1998 a child support grant which is payable to the primary care-givers of children, regardless of their family status. The level of benefit was set at RIOOper month per child for children up to the age of six (incl.). The Department declared that 48% or 3 million children should be targeted. At the same time, the SMGs are to be phased out over a three year period. This research was conducted between November 1995 and March 1998. The analysis of the different suggestions during the policy process and the final policy is based on two pillars: • A situation analysis of the living conditions of South Africa's children on the basis of a composite index. • An evaluation of policy scenarios on the basis of a microsimulation model. The index tries to give a complex picture of the living conditions of children by looking at the financial situation, housing, health, and employment opportunities of the households the children are living in. The analysis reveals that nearly 70% of South Africa's children up to the age of six (incl.) live below the poverty line as defined. A further analysis of the household structure indicates that poorer children are likely to live in larger households. The overall policy shift from a support of single parent families to children in poverty regardless of their family status is espoused. However, the microsimulation model which analyses the impact of different factors like the 'level of benefit', the 'age-cohort', the 'means-test', and the 'administrative requirements', reveals that there are still serious flaws in the current policy. Due to the fact that the means-test is based on the total household income, nearly 40% of the children living below the poverty line are excluded. In addition, the administration needs urgent attention as its capacity is the decisive factor in the success of the programme. The thesis calculates that in the next five years up to R2 billion less will be spent on poor children and the goal of reaching 3 million children will not be achieved, if the problems identified are not addressed. The thesis develops an alternative suggestion to the current policy. While microsimulation has become quite a standard procedure in the analysis of social policies in industrialised countries, there is so far no application in developing countries. It is hoped that by taking this policy analysis as a case-study, this thesis is a step towards the introduction of this method here. Microsimulation models provide important information to enhance the transparency and accountability of policy processes. In this case, civil society was able to challenge Government's decision on a very informed basis, to put pressure on decision makers successfully, and to make workable alternative suggestions. Furthermore, the analysis reveals that against Government's promise redistribution does not take place. Instead a shift towards a more neo-liberal approach in social policy is observed.