Keeping a Foot in the Rural Area: Migration Dynamics, Household Structure and the Link with Socio-Economic Status in Rural South Africa

Mark Collinson, University of the Witwatersrand
Stephen Tollman, University of the Witwatersrand
Kathleen Kahn, University of the Witwatersrand
Martin Wittenburg, University of the Witwatersrand

Background This paper examines the links between socio-economic status, household composition and domestic migration dynamics using data from the annually updated Agincourt demographic surveillance system (DSS). Key questions addressed are: how is the community differentiated by socio-economic status, migration status and employment status; what is the structure of each of these subgroups according to age, gender, educational status, and marital status; how are household types distributed within each group; and to what extent is membership in these groups related to migration behaviour? Importantly, the group that is too poor to migrate is carefully identified, described and contrasted to those that have the means to migrate. The population studied comprises 68,000 people resident in 21villages in the Agincourt sub-district of the Limpopo Province of South Africa. The DSS enables a unique opportunity to look at movement categorisation and the links with other key structural and socio-economic factors. The socio-political structure of the study population at the turn of the new millennium is strongly influenced by former homeland policies and the legacy of landlessness and labour migration. Another important structural feature is the presence of a large group of settled Mozambican refugees. Previous work on population trends have shown rapid fertility decline, an emerging salience of female labour migration, a growing prevalence of smaller household types, and mortality increases associated with AIDS and non-communicable diseases. Methods Migration status is explored at both individual and household levels stratified by movement type. Categories of movement include: non-movement, temporary migration (mostly away but strongly linked to a rural home), and permanent movement (declared permanent intention of residence change). Each category is further differentiated by the extent of movement: internal movement (within the field site), external out-migration (out-migrating from the site), and external in-migration (in-migrating into the site). These definitions are adopted to enable demographic surveillance and involve additional definitions of household membership and the geographic extent of the field site. Further category refinements include marital status and accounting for urbanisation to rural towns or to larger urban centres. Socio-economic differentiation is measured using a composite index at the household level built from an array of variables associated with socio-economic status such as type of dwelling structure, power utilization, water supply and asset ownership. These data were collected in a cross-sectional survey conducted in 2001. The household is defined as the collection of people who usually “eat from the same pot”, or in the case of a temporary migrant, “would eat from” on their return. Household composition is explored through establishing the relationships of co-resident household members to the household head. Households are categorised into various types based on their composition and the types of relationships shared by their members. Data describing household composition and relationships to the household head are updated annually. Employment status is derived from a labour force participation survey conducted in 2000. A locally salient definition of “work” was developed through several workshops with community members and iterative piloting, and previous analysis has shown these data to be sensitive to informal sector occupation. Expected findings The largest group is the group that does not move making up 72% of the population, particularly the elderly and the children. The most sizable category of migration is temporary migration, at 20% of individuals. Temporary migration is still dominated by males (younger and older adults), but the presence of women is expanding rapidly in this group, particularly younger adult women. Temporary migration is the migration category that is most strongly associated with socio-economic status. The poorest group do not temporarily migrate and the likelihood of migrating temporarily increases by about 7% in each successive quintile of the composite SES indicator. Among children there is no gender differentiation in any migration category. Across all children 84% did not migrate in the year of observation; 7% were temporary migrants; 5% were internal migrants; 2% external out-migrants; and 2% external in-migrants. Child temporary migrants were predominantly scholars. Fostering of children can be a reason for migration in all the migration categories.