Labour and migration policy in South Africa

This thesis concerns labour and migration in South Africa. We use a dynamic computable general equilibrium model to analyse the effects of a policy-induced cut to illegal immigration on the local economy. We run the same policy simulation under two distinct modelling scenarios. The first scenario portrays typical labour market conditions and behaviour associated with economies that tend to operate at or near full employment. Simulation results under Scenario 1 indicate that a cut in low-skilled immigration would detract from the long-term welfare of local workers by generating a deterioration in the occupation-mix of their employment. With less low-skilled immigrants, native workers would slip down the skills and earnings ladder. This finding supports modelling evidence for the United States presented in Dixon et al (2011). However, the South African labour market can hardly be described as ‘typical’. The impact of high unemployment and a legal minimum wage at the lower end of South Africa’s labour market stands in contrast to the economy implicitly described under the assumptions of Scenario 1. The second scenario recognises this contradiction between our modelling assumptions in Scenario 1 and available evidence for the South African labour market. A new wage adjustment process assumed for Scenario 2 accounts for the vast surplus of low-skilled legal workers available in the local market. Simulation results under Scenario 2 indicate that a reduction in illegal immigration would benefit legal residents. The favourable result in Scenario 2 relative to that in Scenario 1 is driven by two factors. The first is a much larger gain in legal employment at the lower end of the market. With illegal labour becoming scarcer, employers shift some of their demand for labour towards legal workers. Under Scenario 2, very little wage pressure arises from this increased demand for legal workers, resulting in a significant upsurge in their employment levels relative to that in Scenario 1. The second factor is that the occupation-mix effect is considerably less negative in Scenario 2 than in Scenario 1. In the absence of wage increases in low-skilled occupations, Scenario 1 does not generate a shift in legal work preferences towards these occupations. By presenting both modelling scenarios we are able to better understand and explain our simulation results. We are also able to clearly contrast our set of results against those in the Dixon et al (2011) study of the United States, which used a similar methodology. The importance of applying suitable labour market mechanisms when modelling the South African economy is highlighted. Two main contributions emerge from this thesis. The first relates to the policy simulations. Detailed analysis of the impact of reduced illegal immigration on the South African economy provides policymakers and researchers with fresh evidence based on a state-of-the-art methodology. The second relates to the economic model. In developing the theory and database of the model used in this study we establish a flexible analytical framework to evaluate many other topical issues in South Africa. The detailed and economy-wide nature of the model is well-suited to not only conducting policy analysis, but also forecasting and historical analysis.