Mixed Migration, Forced Displacement A N D Job Outcomes in South Africa E XECUTIVE SUMMARY Mixed Migration, Forced Displacement and Job Outcomes in South Africa 1 Executive Summary S OUTHERN AFRICA HAS A LONG history of human mobility centered around the migration of labor to farms and mines in the region. Patterns of migration and displacement were trans- formed by the end of Apartheid, changing economic systems, and conflict and political instability, both in the region and elsewhere. Today mobility in the region is (i) motivated by a combination of diverse social, political and economic rea- sons, (ii) shaped by long-standing historical movements and reshaped by newer patterns of urbanization and displace- These complex patterns of migration ment, (iii) organized through various legal and extra-legal and displacement, state responses means, and (iv) governed by fragmented and contradic- to them, and the implications for tory legal frameworks. These complex patterns of labor market outcomes in South migration and displacement, state responses to them, and Africa, as the major destination the implications for labor market outcomes in South country in the region, are the subject Africa, as the major destina- tion country in the region, are matter of this study. the subject matter of this study. Scope This study analyzes the characteristics, causes, and conse- quences of migration and forced displacement in Southern Africa. It includes a brief historical overview, as well as an analysis of current migration trends and their impacts. Further, a brief overview of the policy and legal framework governing migration and displacement across Southern Africa is provided Given South Africa’s position as the major destination for migrants and refugees in the region, the primary focus of this study is an analysis of the links between migration and labor market outcomes in South Africa, namely employment and wages between 1996 and 2011. The evidence provided can serve to inform policy in the region. Mixed Migration, Forced Displacement and Job Outcomes in South Africa 1 Mixed Migration and Forced Displacement practices of migration control persist. As a result, patterns of migration and displacement in the region are characterized by into and within Southern Africa elements of both continuity and change. The discovery of diamonds in 1867 and gold in 1886 laid the foundations for South Africa’s highly centralized mining indus- try, and led to the development of the region’s migrant labor TRENDS OF MIGRATION AND DISPLACEMENT IN SOUTHERN AFRICA system. During this period, Southern African economies relied The total migrant population in the Southern African countries on a low-wage, low-skilled, highly-controlled, expendable work- covered in this study (see map below) has increased by about force. Migrant workers, almost all men, were recruited from 68 per cent since 1990. This increase in mobility has not been rural South Africa but also from Lesotho, Malawi, Mozambique, uniform over time, nor across the region: the largest increases Zimbabwe, Zambia and Tanzania, and usually had to make long in migrant stock have occurred in South Africa, Botswana, and hazardous journeys to their workplaces, where they lived Mozambique and Angola, whereas the number of migrants has and worked in dangerous, over-crowded conditions. decreased in the poorer economies of Malawi, Zambia, and The racist regulatory measures which controlled mobility, and Zimbabwe. South Africa is the single largest country of desti- the extent to which they were enforced in different coun- nation. As of mid-2017, UNDESA estimated that it was host tries, had a profound impact on the patterns and distribu- to about 67 per cent. of the regions total migrant population. tion of poverty and inequality in Southern Africa. They also shaped the region’s economies, urbanization, primary (and SOCIO-ECONOMIC CHARACTERISTICS gendered) livelihood strategies as well as forms of political OF MIGRATION IN SOUTH AFRICA leadership, organization, and resistance. In South Africa, for The 2011 Census placed the number of international instance, townships were designed to act as labor reservoirs migrants in South Africa at 2,173,409, about 4.2 per cent. for cities and industries centered around urban spaces. This has had direct implications for settlement patterns in South FIGURE 1:  Total migrant population (both sexes) including refugees African townships, the provision of public services and for the and asylum seekers, in Southern Africa, mid-2017 availability and quality of housing in these townships. Total migrant stock (both sexes) including refugees and asylum seekers, mid-2017 Changes in Southern Africa’s political economy led to a major re-structuring of the migrant labor system across the region. As economies shifted to more capital-intensive forms of growth, unskilled migrant labor became increasingly super- fluous to industry and the need for skilled and semi-skilled labor increased. Angola 638,499 Mozambique Zambia Malawi 246,954 The profound economic and political upheaval surrounding 156,982 237,104 the end of Apartheid in South Africa transformed migration and displacement across the region. Alongside the organized Zimbabwe and controlled labor migration system, other forms of mobil- 403,866 Namibia ity emerged and expanded: this included asylum-seekers flee- Botswana 95,067 ing conflict and persecution, seasonal migrants and cross-border 166,430 traders and smugglers. These changes took place within a broader socio-economic and legal context shaped by the rela- Swaziland tionship between Southern Africa’s system of mine-based capital 33,263 accumulation and the migrant workforce on which it relied. Lesotho 6,749 South Africa The content of laws governing migrants and refugees has 4,036,696 changed in many cases since then, and often radically, but their Total Migrant Stock ≤246,954 implementation remains uneven. At the regional level, migra- ≤638,499 tion continues to be governed by unenforceable conventions 0 200 400 800 Km ≤4,036,696 and a patchwork of bilateral agreements and treaties, while World Bank Geospatial Operations Support Team: GOST (2018) Base Map Source; Esri, USGS, NOAA in many national contexts, bureaucratic and administrative Data Source: United Nations, Department of Economic and Social Affairs. Population Division (2017). Trends in International Migrant Stock: The 2017 revision (United Nations database, POP/DB/MIG/Stock/Rev.2017). 2 Executive Summary of the country’s total population at that time. The major- ity of migrants (around 75.3 per cent.) originated in other One immigrant worker generates African countries, with significant populations from Europe (8.2 per cent.) and Asia (4.7 per cent.). The vast majority of approximately two jobs for locals. all migrants in South Africa, 68 per cent., originated in other SADC Countries. Mixed Migration, Forced Displacement Not all provinces in South Africa attract migrants to the and Job Outcomes in South Africa same extent. Gauteng, in particular, as the primary economic Migrants and refugees in South Africa are often entangled in and financial hub in South Africa, hosted an overwhelming political discourse that blames them for ‘stealing’ local jobs. majority of international migrants with 52 per cent. of the The implications of migration on local jobs, is therefore, total. South Africa is also among the more highly urbanized highly contested. However, empirical evidence may provide countries in Africa, and Gauteng is particularly highly urban- the necessary information policy-makers require to develop ized. As a result, migration and displacement in South Africa policies and interventions that mitigate the costs that may be has important urban dimensions. felt by locals, while enhancing the developmental opportuni- According to the same 2011 Census, 39.8 per cent., of inter- ties for migrants, refugees, locals and the wider economy. national migrants in South Africa were women. Interestingly, It is within this context that this study estimates the impact however, among migrants from SADC, more women than of immigration on labor market outcomes such as employ- men aged 15-24 had migrated to South Africa (reversing his- ment and wages in South Africa between 1996 and 2011, torical trends). Around 23 per cent. of migrants fell below the and posits several possible explanations for what might national poverty income level, and in general, over half of all account for the results. international migrants in 2011 were poor. Around 27 per cent. of all households were headed by women, but almost one- There are only a few papers that have studied the impact of third of these female-headed households fell in the national immigration on labor market outcomes in South Africa. They poverty category. Around 63 per cent. (or three out of five) have generally found no impact on total income but nega- international migrants reported being employed. The major- tive effects on local employment. Compared to earlier papers, ity were employed in the formal sector, while 17.2 per cent. this analysis makes several contributions. First, the analysis reported earning their livelihood in the informal sector. 17 uses industry-province level data, given significant variation per cent. were employed in private households. in the utilization of immigrant labor across industries and Mixed Migration, Forced Displacement and Job Outcomes in South Africa 3 provinces. Applying data on a more aggregate level such as OLS results may be affected by simultaneity bias, the analysis has provinces instead of smaller geographical units such as districts also produced IV estimates for Equations (1) and (2) (See chap- reduces effects from potential outflows of locals in response ter 5), using the instrument described and its square to permit to immigrants’ inflows. Second, the analysis uses an instru- overidentification of the equation. In both forms of the model, mental variables (IV) approach to address endogeneity issues. the immigrant growth rate for the industry, province and year The instruments are constructed following the methodology has significant and positive effects on total earnings and wage proposed by Card (2001), who uses previous settlements of earnings with values ranging from 0.3 to 1.4. immigrants as an instrument in studying labor market effects of immigration across geographical regions. Third, the analy- UNDERSTANDING THE RESULTS sis includes all immigrants–not only males, as in some stud- This analysis suggests several explanations for the positive ies–given a substantial share of female employment among impact of immigrants on South Africa’s labor market. First, immigrants. In 2011, females accounted for nearly 24 percent given that immigrants and locals are not perfect substitutes, of total employment among immigrants. Next, the estimates specialization in different tasks might lead to overall produc- focus on the relationship between locals and immigrants, and tivity gains. Second, immigrants have nearly twice as high an not on specific groups based on education and experience employment-population ratio compared to locals, possibly within each category. The latter captures only the partial reflecting the demand for the diverse set of skills they bring own-skill effect and ignores cross-skill complementarities and and this can result in large multiplier effects. Finally, immi- externalities. Finally, the study uses wage data from the Post- grants tend to be more risk-taking and entrepreneurial, which Apartheid Labor Market Series (PALMS) harmonized survey, might generate positive externalities in the economy. instead of relying on total income that includes both labor and non-labor earnings as in other studies. For details on the Foreign and local-born workers might specialize in perform- methodological approach, see chapter 4. ing complementary tasks and, hence, the two groups might not compete for similar jobs. As a result, this complemen- RESULTS tarity might increase the productivity of local workers and The estimation results indicate that immigration has a positive hence generate positive externalities for their employment impact on local employment, labor earnings, and wages. The and wages. estimated effects of immigrant growth on local employment are To test this hypothesis the study compares occupations of positive and highly significant in all specifications and are similar immigrants and locals for tertiary and non-tertiary educated in terms of magnitudes. They show that a one percent increase groups. It uses 2011 census data and constructs the Welch in the number of immigrants relative to the previous period (1979) index to test whether immigrants and locals are per- raises local employment by 0.2 percent. In other words, one fect substitutes. This index is similar to a correlation coeffi- immigrant worker generates approximately two jobs for locals. cient, equaling one (1) when the two groups have identical The effects on labor earnings, wages, and self-employment earn- occupational distributions and minus one (-1) when the two ings are insignificant in OLS estimations with an exception of groups are clustered in completely different occupations. We wage earnings when only individual fixed effects are used.As the obtain .35 for the non-tertiary educated group and .14 for the  mployment-Population Ratio FIGURE 2: E  hare of Self-employed in Total Employment FIGURE 3: S IMMIG RANTS NATIVES IMMIGRANTS N AT I V E S 0.80 0.30 0.25 0.61 0.60 0.55 0.20 0.18 0.16 0.40 0.36 0.35 0.35 0.31 0.10 0.10 0.08 0.20 0.06 0.00 0.00 1996 2001 2011 1996 2001 2011 4 Executive Summary tertiary educated group.This shows that immigrants and locals therefore these results may differ in the current context. It are not perfect substitutes and the complementarity of tasks is also well documented that even in the best circumstances, they perform might generate efficiency gains in the economy. migration and displacement may have significant short-term costs for receiving communities. Finally, two stylized facts emerge from the comparison of the employment-population ratio and share of self-employed in total employment between locals and immigrants. First, Conclusion although the employment-population ratios for locals and Labor Market Outcomes: An Opportunity immigrants were similar in 1996, this drastically changed by for Policy Dialogue and Further Research 2011 (Figures 2 and 3). In particular, the share of employed The results and substantiations provided here, are significant immigrants in total increased from 36 to 61 per cent. com- for policy makers and development actors in South Africa pared to only 35 per cent. for locals. and the wider region, and as such, their implications should This sharp increase in the employment-population ratio for be seriously considered. They provide a basis for substantive immigrants was probably due to changes in immigration pol- policy dialogue on how to enhance the development impacts icies post-Apartheid. Higher employment rates among immi- of migration, especially for local job and wage outcomes and grants and hence higher labor earnings compared to locals the South African economy. Critically, although such quanti- might generate large multiplier effects in the economy. tative analysis is instructive, perceptions and subjective evalu- ations of well-being of both locals and migrants, matters. As Finally, we must also note the prevalence of self-employment such, any interventions in response to these results, should among immigrants: self-employment accounted for 25 per- account for perceptions and lived experiences. Equally, the cent of total jobs for immigrants, compared to 16 percent for political will to advance policies based on empirical evidence, locals. Migrants are more likely to appear in entrepreneurial is a necessity for the achievement of any sustainable and posi- roles than locals, suggesting that their actions are likely to pro- tive economic outcomes for locals and migrants, alike. mote economic growth by enhancing, for instance, the supply of small retail establishments. If those businesses are success- Crucially, these results also provide an important foundation ful, they also will provide multiplier effects which may spread upon which further large-scale research can be developed. beyond the immediate family. Such research can potentially complicate and enable a richer understanding for how migrants, refugees and asylum-seekers An important note of caution is that these results are retro- engage with the South African economy and the economic spective in nature given the data limitations mentioned, and relationships they share with locals. This research, may for instance, consider how circular migration, informality, undoc- umented movements and gendered dimensions have implica- tions for economic relationships. The results and substantiations provided here are significant for policy makers and development actors in South Africa and the wider region, and as such, their implications should be seriously considered. Mixed Migration, Forced Displacement and Job Outcomes in South Africa 5 “One immigrant worker generates approximately two jobs for South Africans” — World Bank, 2018 For additional information, please contact: Helidah R.A. Ogude Social Development | Africa Region hogude@worldbank.org 4 Annexures