South Africa has a history of distorted and controlled migration. Remnants of this
history are still present to this day. The purpose of this study is to understand the
patterns of migration in contemporary South Africa. In particular we focus on the
interactions between migration and labour force participation decisions. Using the
GPS coordinates in South Africa’s first nationally representative panel dataset,
the National Income Dynamics Study (NIDS), migration is defined as a movement
of individuals across municipal boundaries between waves of the NIDS survey.
The analysis then goes on to explore the factors driving this migration. A range
of relevant individual and household variables are available in NIDS. In addition
community level factors such as socio-economic indicators and local service delivery
are derived from Census and Community Survey and merged into NIDS in order
to provide a rich dataset. Descriptive analysis is followed by the estimation of
a biprobit model of migration and participation. Thereafter, the post-migration
earnings of migrants are estimated while accounting for selection. The young,
educated and the relatively better-o! in migrant communities are more likely to
migrate and individuals are found to migrate out of communities with high levels
of relative inequality. The interdependence of the migration and participation
decisions is a"rmed. In modeling earnings of migrants we find that the selection
into migration has a negative e!ect on wages especially for high income earners. In
general we find that South Africa is beginning to report similar trends in migration
to its developing country peers.