This study is therefore a first attempt to analyze the impact of internal in-migration on income inequality of receiving areas and is placed in the context of South Africa. The issue is relevant to no other country as much as to South Africa, which has one of the highest income inequalities in the world attributable to historical factors such as the discriminatory policies of the Apartheid regime (1948-1994) against women and non-whites. Despite the democratic government of South Africa instituting a series of legislations meant to ensure wage equality across race groups and genders, income inequality has been increasing post-democracy from a gini coefficient of 0.59 in 1993 to 0.65 in 2011 (World Bank 2013).
The issue of migration is also of specific interest in the South African context. Under the Apartheid regime, the movement of the vast majority of the population was restricted through the oppressive Group Areas Act and Influx control policies (Zuma, 2013). The elimination of these policies meant that in the Post-Apartheid years the country experienced accelerated urban migration (Mulcahy and Kollamparambil 2016). The intersection of inequality with migration arises from the finding of Leibbrandt, Finn and Woolard (2010) that while urban inequality has increased since 1993, rural inequality seems to have fallen. This points to the possibility of rural-urban migration playing a role in the emerging trends in inequality. Exploring the role of in-migration in explaining this phenomenon of rising urban inequality is hence of high relevance in the context of South Africa.
The study which is undertaken at the district level makes use of the National Income Dynamics Survey data to map the movement of individuals between districts over the period 2008-2015. The results of our multivariate regression analysis using system GMM technique that enables the inclusion of past inequality as well as effectively counters the reverse causality issues, validate the persistent nature of regional income inequality. A 1 % increase in in-migration into a district increases individual income inequality by 0.02%. This result is statistically significant at 99 percent significance level. The result is unsurprising because migrants constitute predominantly of less-educated workers who are likely to find employment in the informal sectors, thus increasing the level of inequality in the receiving areas. The level of employment is also seen in our estimation models to be critical at 99% confidence level in determining the inequality levels. Increased employment rate will effectively reduce inequality by 0.11%. Another relationship that comes out consistently across the models is that districts with higher levels of average income has lower inequality. Districts with higher proportion of population with education level above matriculation is seen to have higher levels of inequality at 95% confidence level.
Our analysis shows that rising urban inequality in the urban areas as indicated by Leibbrandt, Finn and Woolard (2010) can be attributed at least in part to rural-urban migration. This works through both the wage as well as employment channel. The employment channel can be said to have a stronger impact than the wage channel as indicated by the coefficients estimated through our multivariate regression analysis. The higher rates of unemployment among the migrants as compared to the non-migrants therefore becomes a cause of worry. The policy implications of the study are clear in highlighting the need to address unemployment issues in general, and among migrants in particular, in making a dent on inequality. Improving the employability of migrants is critical in reducing the inequality, therefore interventions to reduce urban income inequality cannot ignore education and capacity building within rural areas. Addressing education and skill formation in rural sector will improve the quality of migrants that can be absorbed in the urban formal sector without increasing inequality. In conclusion, it can be said that the urban bias in policy formulation that ignores the rural sector cannot be successful as it can only lead to the opposite effect as indicated by the Harris Todaro model.
One limitation of the analysis is the study period allowed only the assessment of impact in the short-run. The longer run impact of in-migration can differ from the short-run impact and needs to be studied further within a longer time frame.