Nalen Naidoo, 1 Murray Leibbrandt 2 and Rob Dorrington 3

Transcription

1 SADemJ (11) Magnitudes, Personal Characteristics and Activities of Eastern Cape Migrants: A Comparison with Other Migrants and with Non-migrants using Data from the 1996 and 2001 Censuses Nalen Naidoo, 1 Murray Leibbrandt 2 and Rob Dorrington 3 Abstract This paper investigates the changing nature of migration of the African population from the former Transkei, particularly the rural to urban migration to the Cape Metropolitan Area over the period 1991 to 2001 using data from the 1996 and 2001 censuses. The study compares the characteristics of those who migrate from the Transkei to the CMA with those who migrate within the Eastern Cape and those who do not migrate and investigates whether the characteristics of these migrants have changed significantly over this time. Nationally there has been an increase in migration of females and the young and of migration to non-metropolitan areas. Migration flows appear to have stabilised. However this is not uniformly the case with migration from the Eastern Cape to the Western Cape remaining strong. The labour market environment confronting migrants has worsened with significant increases in percentages of migrants who are unemployed and some de-skilling of the occupations in which migrants are finding employment. The dire labour market situation in rural Eastern Cape maintains the flow of migrants despite this hostile environment for migrants. INTRODUCTION In South Africa, as is the case in many countries, internal migration is a social force with profound demographic, political and economic implications. Yet, as in other countries, it is notoriously under-researched (Fix 1999:7). This is 1 Group Financial Reporting, Aviva plc, London, United Kingdom 2 School of Economics, University of Cape Town, South Africa (Corresponding author) 3 Centre for Actuarial Research, University of Cape Town, South Africa 3

2 4 Southern African Journal of Demography 11(1) so, not because of a perceived lack of importance but, rather, the intrinsic difficulty of measuring migration (Fix 1999:7) and the poor quality and inappropriateness of data available on the topic. During apartheid, much research focused on the movement of the African population under policies designed to channel and encourage this movement to certain areas and stem the flow to others. The result of these activities was the creation of a reservoir of migrant labour, located in such a way as to make their labour accessible, but kept relatively far away from urban and metropolitan areas. This has created a large group of people who, in the post-apartheid era, still reflect the effects of policies of the past but who are prepared to move to redress the imbalances of their circumstances. Apartheid policies laid the conditions for a migratory population that had to seek out alternative places of residence, however temporary, in an attempt to improve the poor quality of life. In the words of one African worker, The countryside is pushing you into the cities to survive, and the cities are pushing you into the countryside to die (Savage 1984:50). It is hardly surprising that a review by Posel (2003) finds that research in the 1970s and 1980s focused on the characteristics of this migrant labour system. More puzzling is the fact that, since 1986 (with the abolition of influx control), empirical research on internal migration has dwindled and, in the 1990s, the focus seems to have moved mainly to immigration issues. Posel (2003) surmises that this may be due to an assumption that temporary and circulatory labour migration has been replaced by permanent migration in post-apartheid South Africa. The majority of migrants out of the Eastern Cape originate in the former Transkei. The Transkei is one of the largest and poorest of the former homelands in South Africa. As such, it is of special interest in discerning some of the post-apartheid responses of those who bore the brunt of what Savage (1984:3) has called the disorganisation and reorganisation of the African population under apartheid. The region is the area of origin of the majority of the African migrants entering the Western Cape and the profiles of these migrants and the routes they take have been the subjects of many studies (Bekker 2002; Cross and Webb 1999). According to Bekker (2002:10) this stream is currently the largest and most rapid demographic flow in South Africa. Leibbrandt et al. (2002) used the 1996 census micro data to analyse the factors affecting the migratory tendencies of the African population in the former Transkei, focusing particularly on African rural-urban transitions to the Cape Metropolitan Area (CMA). The present study extends this research,

3 Magnitudes, Characteristics and Activities of EC Migrants 5 using both the 1996 and 2001 censuses to profile migration out of the former Transkei for the periods 1991 to 1996 and 1996 to Through the development of this profile, a better understanding of the people leaving the Eastern Cape and entering the Western Cape can be gleaned. Based on this understanding, government policies could be designed specifically to meet the needs of these migrants. The analysis makes use of the Census 1996 and Census 2001 Community Profile Databases, and the Census 1996 and Census % Samples (Statistics South Africa 1998 and 2004). The paper proceeds as follows. The following section begins by listing some of the strengths and weaknesses of the census data in addressing migration. The heart of the paper is the presentation of a substantial amount of descriptive information on the migration flows and on the migrants, in the third section. This section starts by comparing the magnitudes of migration out of and within the Eastern Cape with other national migration flows between rural and urban areas and between provinces. It then describes the characteristics of Africans migrating out of the Eastern Cape and compares these characteristics to nonmigrants. Profiles are presented by race, age, employment and occupation and education. The concluding section of the paper then uses this descriptive platform as a base from which to draw out some implications for the South African migration literature and for policy towards migrants. Using Data from the Census to Analyse Migration The 1996 and 2001 censuses were conducted in October of those years. New municipal boundaries were implemented in 2000, and new demarcations were made that divided provinces into district councils and metropolitan areas. Despite the complications this introduced, there is a general consistency between the 2001 and 1996 censuses (Statistics Council, 2003). However, there are a number of comparability issues that need to be mentioned. Most importantly, it is ironic that prior to Census 1996, the censuses generally were not representative of the population in South Africa but contained detailed questions on migration while after that the census became more representative but the questions on migration progressively less detailed. The 1996 census, for example, asked respondents to quantify remittances from migrants, while the 2001 census omitted all remittance questions, as well as questions that related directly to migrant labour. The migration-related questions asked in the census are also problematic. Circulatory and oscillatory migration are impossible to measure, and regionalisation of the household

4 6 Southern African Journal of Demography 11(1) (Bekker 1999), where family ties are so close that households in rural and urban areas operate as one, serves to confuse questions regarding place of usual residence. Also, migration within a main place could not be analysed. Then, as data are only available as at the time of the census, it is not possible to look at changes to the situation of migrants before and after the migration. For example, changes in employment status or income level due to a move could not be ascertained (Leibbrandt et al. 2002). Finally, as regards the migration variable, throughout this paper a migrant is defined as a person who moved suburb, ward, village, farm or informal settlement at least once during the five-year period prior to the date of the census. This analysis excludes children born during these respective five-year periods. This is because a migrant would have effectively answered No to the question: Were you living in your current residence at the time of the previous census? Children born between censuses would, thus, fall outside the scope of this question and obtaining the details of any move they may have made would add another level of complexity, apart from having to deal with complications of under-enumeration and scanning error (Statistics Council 2003). There are a few other data issues that are relevant to the analysis of migration. First, there is the definition of urban and rural areas. In the 1996 census, an urban area was defined as an area that fell within a municipality or local authority, and included an ordinary town or city with formal structures, informal dwellings, mining hostels and hospital and prison institutions (Statistics South Africa 2003). Census 2001, while describing an urban settlement as structured and organised, with formally planned and maintained roads and the provision of services such as water, electricity and refuse removal, did not classify enumeration areas into urban and rural (Statistics South Africa 2003). However, for the purposes of comparability between the censuses, the 10% sample of the 2001 data was coded to include the classification of urban and rural according to the 1996 definition. For the first time in South Africa, the 2001 census used imputation to replace unavailable, unknown, incorrect or inconsistent responses. This procedure involves replacing a response (non-blank) or non-response (blank) with a value determined either logically from other responses, or, where this is not possible, from a hot deck. (Hot decking entails using the response from the most recently processed response from a person or household that was similar to the person/household with the missing response.) Table 1 shows the extent to which variables used in this paper were subjected to imputation.

5 Magnitudes, Characteristics and Activities of EC Migrants 7 Table 1 Percentage of each type of imputation undertaken for each variable of interest Variables No imputation Logical imputation (from blank) Logical imputation (non blank) Hot deck imputation (from blank) Hot deck imputation (non blank) Figures shown as percentages Age Gender none Marital status Previous residence None none Year moved None none Province of previous residence None none Main place of previous Level of education Source Derived from the Census % Sample, Statistics South Africa (2004) Table 1 shows that levels of imputation ranged from a high of over 23 per cent for Age to a low of little over 1 per cent for Year moved and Gender. In most cases it was under 6 per cent. The two variables with the highest levels of imputation were: Level of education, which had the highest level of imputation, with a relatively high proportion being hot-deck from blank and logical from non-blank. Age, in which just under a quarter of all observations were corrected. The majority of the corrections were logical imputations from non-blank, which were mostly corrections to age based on date of birth. While the level of imputation is generally higher than the 2 per cent level targeted, 4 the intention of this process was to improve the overall quality of the data. Inspection suggests that there is little difference between the distributions of each variable including or excluding the imputed data and thus the imputed data are used. Two final points about the data are worth noting. First, the unemployment levels that are used in the analysis in the paper were calculated from the census. It is known that these data produce higher unemployment levels than the official statistics for that time (Statistics South Africa 2004:53). 4 Personal communication with Professor Jacky Galpin, Chair of the 2001 Census Sub-Committee of the South African Statistics Council.

6 8 Southern African Journal of Demography 11(1) Second, this study will not include any analysis of the income of migrants. This is largely due to the problematic nature of the income variable in the census, with widespread under-reporting, and the high level of imputation undertaken for the variable (Cronje and Budlender 2004). A PROFILE OF MIGRANTS: and This section presents a general description of the characteristics of migrants in South Africa, based on the 1996 and 2001 Census Community Profile databases, and the 10% Samples (Statistics South Africa 1998 and 2004). All results using the 10% Samples are calculated using the person sample weights supplied by Statistics South Africa. After considering migration at the national level, we examine the characteristics of African migrants compared to African non-migrants and non-african migrants from the Eastern Cape to the Western Cape, from the Eastern Cape to the rest of South Africa, and migrants remaining within the Eastern Cape. The results from the 2001 census are contrasted with those from the previous census. This comparison assumes that the characteristics of Eastern Cape migrants approximate those of migrants from the former Transkei. Figure 1 Proportion of the South African population in urban areas 70% 60% 50% Proportion 40% 30% 20% Total South African Total African 10% 0% Year Source HSRC (1996): Socio-economic Atlas for South Africa, African; Savage (1984:20); Census 1996 and % Sample, Statistics South Africa (1998 and 2004) 5 5 The data for Total South African were sourced from Socio-economic Atlas for South Africa and the Census % Sample. The data for Total African were sourced from Savage (1984), and the Census 1996 and Census % Samples.

7 Magnitudes, Characteristics and Activities of EC Migrants 9 Urbanisation We begin with an analysis of movements to urban and metropolitan areas. As Figure 1 shows, there has been a strong trend of urbanisation in South Africa with just over half of the South African population residing in urban areas by This trend accords with Kok et al. (2003) who noted a similar strong trend to However, this level of urbanisation still falls short of the types of levels predicted by Todaro (1976) under the mobility transition hypothesis. According to Esau et al. (2004), this is largely due to: the existence of influx control laws that slowed urbanisation until about 1986, circular migration, that has been found to feature in migration patterns, and a rising incidence of migration between informal settlements (intra-urban migration). In addition, the stagnation in growth of employment opportunities in urban areas has probably played a part, too. Table 2 Percentages of the population in urban areas Percentage Population countrywide African population countrywide Non-African population in the Western Cape African population in the Western Cape Non-African population in the Eastern Cape African population in the Eastern Cape As at Census As at Census % increase Notes The 1996 definition of an urban area is used for the calculations in this table. Source Census % Sample, Statistics South Africa (1998) and Census % Sample, Statistics South Africa (2004) As can be seen from Table 2 and Figure 1, despite the lower proportions relative to the whole population, there has been steady urbanisation of the African population with nearly half living in urban areas by 2001 (an increase of 9 per cent from 1996). In contrast, only 31 per cent of Africans in the Eastern Cape were living in urban areas in 2001, with most still living in

8 10 Southern African Journal of Demography 11(1) the former homelands in this province. This is despite the fact that the increase in urbanisation amongst this group (8 per cent) was greater than for the general population (5 per cent). On the other hand nearly 95 per cent of Africans in the Western Cape were living in urban areas, although most African adults in these areas are rural-born (Bekker 2002). Part of this trend is the strong flow of migrants to South Africa s major metropolitan areas (metros). These areas are generally large economic hubs and migrants have been found to be willing to travel long distances to these large centres due to the range of products, services and opportunities on offer. Table 3 shows the proportions of migrants whose destinations were various metropolitan and non-metropolitan areas (including those who moved from one metropolitan area to another, or one non-metropolitan area to another). Kok et al. (2003) defined two different types of metropolisation ; primary metropolisation, which involves rural to metropolitan movement, and secondary metropolisation, which is either a town to metro or a city to metro movement. Tables 2 and 3 do not distinguish between these two types of metropolisation. In the 2001 census, three Gauteng metropolitan areas were defined, whereas in the 1996 census, all were combined into one. The six metropolitan areas in Census 2001 (as demarcated by Statistics South Africa) are the destination areas of 48 per cent of all migrants, with the remaining 52 per cent moving to the more numerous non-metropolitan areas. There has been an increase in Table 3 Percentages of migrants who moved to metropolitan or non-metropolitan areas in the previous five years All African Non-African All African Non-African Cape Town Durban Gauteng East Rand Johannesburg Pretoria Port Elizabeth Non-metropolitan Source Census % Sample, Statistics South Africa (1998) and Census % Sample, Statistics South Africa (2004)

9 Magnitudes, Characteristics and Activities of EC Migrants 11 in metropolisation 6 from the 43 per cent who had moved to metropolitan areas by the time of Census However, the strong non-metropolitan migration provides some support for the claim by Cross and Webb (1999) that there has been a strong trend towards peri-urban settlement and a densification around small towns and cities. They suggest that this may be due to the difficulties of travelling to metropolitan areas, with these small urban centres offering the only feasible contact for the rural population with the urban economy. Johannesburg, East Rand and Pretoria can be grouped into the Gauteng metropolitan areas due to their proximity and to ensure consistency of comparison with the 1996 census. This is the largest receiving metropolitan area, receiving a full 28 per cent of all migrants. Cape Town is second, receiving 9 per cent to 10 per cent of all migrants. It is interesting to note that the majority of non-african migrants move to metropolitan areas (53 per cent in the 1996 census and 61 per cent in the 2001 census) while the majority of African migrants (64 per cent in Census 1996 and 57 per cent in Census 2001) move to non-metropolitan areas. This may be due to the fact that non-africans are more likely to have the necessary social networks in place in metropolitan areas, making the transition to these areas easier. However, the increase in the proportion of Africans migrating to metropolitan areas may indicate that urban migration is becoming easier for potential African migrants. Inter-provincial Migration Table 4 shows inter-provincial migration in South Africa for the period 1991 to 1996 and Table 5 shows this migration for the period 1996 to During the former period, 2.6 per cent of the total population born at least five years prior to the censuses was involved in these migratory flows, while the corresponding proportion for the latter period was 2.8 per cent. Gauteng was clearly the main destination province for both periods ( and , respectively), followed by the Western Cape ( and ). The Eastern Cape was by far the largest sending area ( and ), followed by Gauteng ( and ) and Limpopo ( and ). 6 For these purposes, metropolisation is defined as a process where a migrant chooses a metropolitan area over a non-metropolitan area as destination. 7 It should be noted that the equivalent figure according to Kok et al. (2003) is 75 per cent. It is not clear how they arrived at this figure but they must have used a different definition of metropolitan area.

10 12 Southern African Journal of Demography 11(1) Table 4 Inter-provincial migration in South Africa Province of origin Province of destination EC FS GT KZN LP Eastern Cape (EC) 16,375 54,474 31,868 1,193 Free State (FS) 3,337 31,922 3, Gauteng (GT) 11,729 24,000 24,634 18,081 KwaZulu-Natal (KZN) 5,813 6,671 73, Limpopo (LP) 489 2, ,670 1,502 Mpumalanga (MP) 878 3,301 55,781 4,763 8,829 Northern Cape (NC) 1,674 5,162 6, North West (NW) 785 8,445 94,437 1,304 4,685 Western Cape (WC) 10,100 3,233 13,741 4, Total (SA) 34,805 69, ,156 72,410 35,313 Source Census 1996 Migration Community Profile, Statistics South Africa (1998) Table 5 Inter-provincial migration in South Africa Province of origin Province of destination EC FS GT KZN LP Eastern Cape (EC) 16,810 90,032 59,729 6,368 Free State (FS) 8,761 60,031 8,556 4,380 Gauteng (GT) 29,166 25,205 45,003 39,652 KwaZulu-Natal (KZN) 18,233 8, ,948 7,065 Limpopo (LP) 2,679 4, ,142 5,094 Mpumalanga (MP) 3,187 5,720 88,950 11,249 18,143 Northern Cape (NC) 2,954 7,635 11,060 1,850 1,719 North West (NW) 4,302 10, ,719 4,352 11,602 Western Cape (WC) 26,688 5,235 32,602 9,314 2,491 Total (SA) 95,970 67, ,452 85,418 85,052 Source Census 2001 Migration Community Profile, Statistics South Africa (2003) The Eastern Cape has the Western Cape as its major receiving area, with 41 per cent of its migrants moving there over the period This is slightly less than the 45 per cent for the previous five years. Of the small flow of people leaving the Western Cape, 40 per cent moved to the Eastern Cape over the period It is probable that much of this is return migration of people previously from the Eastern Cape. This flow has increased since the period, when only 25 per cent of migrants out of the

11 Magnitudes, Characteristics and Activities of EC Migrants 13 Province of destination Province of origin MP NC NW WC Total 6,222 1,746 10, , ,314 Eastern Cape (EC) 4,912 4,905 15,556 6,218 71,205 Free State (FS) 38,711 3,777 40,782 35, ,966 Gauteng (GT) 11, ,490 10, ,782 KwaZulu-Natal (KZN) 29, ,969 1, ,202 Limpopo (LP) 554 5,553 1,963 81,622 Mpumalanga (MP) 980 5,184 14,172 34,327 Northern Cape (NC) 4,270 9,238 2, ,569 North West (NW) 2,011 5,465 1,434 40,546 Western Cape (WC) 98,320 26,743 97, ,965 1,048,533 Total (SA) Province of destination Province of origin MP NC NW WC Total 10,087 4,142 21, , ,761 Eastern Cape (EC) 6,991 6,417 20,119 13, ,511 Free State (FS) 34,721 6,829 53,413 58, ,992 Gauteng (GT) 18,852 1,893 7,910 24, ,247 KwaZulu-Natal (KZN) 37,739 1,385 21,374 5, ,074 Limpopo (LP) 1,486 11,560 6, ,111 Mpumalanga (MP) 1,429 7,529 21,430 52,652 Northern Cape (NC) 6,354 16,360 7, ,771 North West (NW) 3,133 9,649 3,769 66,193 Western Cape (WC) 109,219 44, , ,514 1,257,551 Total (SA) Western Cape moved to the Eastern Cape. It is also interesting to note that, over both five-year periods, Gauteng was both a major receiving and a major sending area. It seems that major flows emanate from provinces where the former homelands were located (Eastern Cape, Kwa Zulu-Natal, Limpopo and North West), towards provinces that are economically relatively better off, and have large metros (Gauteng and the Western Cape). It can, thus, be assumed that

12 14 Southern African Journal of Demography 11(1) the processes of urbanisation and metropolisation mentioned above are driving these flows. Of course these proportions are not the same for all ages and, as will be shown below, the bulk of the urbanisation and metropolisation occurs in the young adult working ages, whereas the other migration streams will be concentrated at different ages, including some return migration from urban areas at around retirement ages. Figure 2 and the Tables A.1 and A.2 in the Appendix record the net interprovincial in- migration. It can be seen that Gauteng and the Western Cape have received the largest overall numbers due to migration, over both periods. Every other province (except Mpumalanga) experienced a net loss during both periods with Limpopo and the Eastern Cape showing the highest net losses. Aside from a small net loss to Mpumalanga over the period , the Western Cape is also the only province with a net gain from each of the other provinces during both periods. The Eastern Cape is the only province with a net loss to each of the other provinces, over both periods. Gauteng, though dominating the migration flows, has net outflows to the Western Cape over both periods ( for and for ). The next section considers the profile of the migrants themselves. Figure 2 Net inter-provincial migration in South Africa Migration gains or losses (number of people) Gauteng Western Cape Mpumalanga Free State Northern Cape North West KwaZulu- Natal Limpopo Eastern Cape Province Source Census 1996 and Census 2001 Migration Community Profiles, Statistics South Africa (1998 and 2003)

13 Magnitudes, Characteristics and Activities of EC Migrants 15 The Characteristics of Migrants Racial composition Table 6 shows that 12 per cent of the South African population moved from one main place to at least one other place in the five years prior to Census Compared with the five years prior to the 1996 census, it seems that the population is becoming more settled in its residential areas, as about 20 per cent of people were involved in moves between 1991 and However, given the increase in inter-provincial migration, it seems that those who do move were more likely to change provinces. The White population seems much more mobile than any other group with 41 per cent in 1996 and 27 per cent in 2001 moving between main places. Bekker (2002) surmises that this may be a result of Whites having superior access to information, a housing market and financial resources, and a lower reliance on family support networks. This migration includes a tendency for Whites to move to the Western Cape when they retire. The decrease in proportions migrating could indicate that the White population is becoming more settled, or that the job and resettlement opportunities for this group have become more limited. Table 6 Migrants between main places by population group Population Group 1996 Population (community profile) Migrants ( ) Migrants as a proportion of the 1996 Census population (%) 2001 Population (community profile) Migrants ( ) Migrants as a proportion of the 2001 Census population (%) Black African Coloured Indian White Total Source Census 1996 and Census 2001 Migration Community Profiles, Statistics South Africa (1998 and 2003) The general level of movement amongst Africans is proportionately much lower than other groups (17 per cent and 11 per cent in each period). These lower relative proportions and their decrease over time may indicate that 8 Movement within main places is not considered in this calculation.

14 16 Southern African Journal of Demography 11(1) movement amongst the African population is stabilising at its current levels, after the volatility of the pre-1990s. Bekker (2002) also noticed a fall off in migration amongst Africans (to the Western Cape in particular). Table 7 Percentages of migrants split by population group Population Group (%) countrywide Migrating countrywide in the EC Migrating from the EC to the WC Migrating from the EC to the rest of SA Migrating within the EC Black African Coloured Indian or Asian White Black African Coloured Indian or Asian White Source Census 1996 and Census % Sample, Statistics South Africa (1998 and 2004) Table 7 offers a further analysis of the ethnic composition of these migrants over the and periods. It shows that, whereas only 68 per cent of the countrywide migratory population is African in the latter period (up from 64 per cent in the 1996 period), 86 per cent of the stream from the Eastern Cape to the Western Cape, and 79 per cent of the stream from the Eastern Cape to the rest of South Africa are African (down from 85 per cent and 88 per cent, respectively, in the 1996 period). This accords with the largely African composition of the Eastern Cape. Of note, is the decrease in the proportion of Africans migrating from the Eastern Cape to the rest of South Africa excluding the Western Cape (88 per cent in the 1996 period to 79 per cent in the 2001 period), and the large increase in the proportion of Coloureds in this stream (2 per cent to 8 per cent). Age Figure 3 shows the age distribution of migrants. Additional detail is provided in Figures A.1 and A.2 in the Appendix. We see that about 59 per cent of all people moving from the Eastern Cape to the Western Cape between 1996

15 Magnitudes, Characteristics and Activities of EC Migrants 17 and 2001 fall into the 20 to 39 age band. The equivalent proportion for was 56 per cent. The 20 to 24 age group represents the largest proportion moving from the Eastern Cape during both periods. With respect to non-migrating Africans in the Eastern Cape, about 34 per cent and 32 per cent fall into the 5 to 14 year age band for and respectively, while 16 per cent fall into the 50 year and above age band over both periods. This effectively means that, since the early 1990s, almost half of the population remaining in the Eastern Cape falls outside of the prime working ages. Of the streams originating from the Eastern Cape, it appears from Figures 4 and 5 that the stream flowing to the Western Cape is somewhat more youthful than the stream flowing to other parts of the country. Also noteworthy is that about 20 per cent of migrants within the Eastern Cape (for both five-year periods) are in the 5 to 14 year age band. While this situation may reflect the age structure of the Eastern Cape population and movement in these ages could be connected to changes in schooling, it is also possible that these may be children having to move between different support bases while they wait to follow their highly mobile parents, who may have already left the province in search of work. The migrants within the Eastern Cape also appear older than those leaving the province. Figure 3 Age distribution of migrants 100% Proportion 90% 80% 70% 60% 50% 40% 30% 20% 10% 50 years and above 45 to 49 years 40 to 44 years 35 to 39 years 30 to 34 years 25 to 29 years 20 to 24 years 15 to 19 years 5 to 14 years 0% African Countryw ide Migrating Non-African Countryw ide Migrating African Countryw ide A frican in the EC Stream Migrating African from the EC to the WC Migrating African from the EC to the rest of SA Migrating African within the EC Source Census 1996 and Census % Samples (Statistics South Africa, 1998 and 2004)

16 18 Southern African Journal of Demography 11(1) Figure 4 Age profile of African migrants originating in the Eastern Cape, % Proportion 35% 30% 25% 20% 15% 10% 5% African in the EC from the EC to the W C from the EC to the rest of SA within the EC 0% Age Grouping (years) Source Census % Sample, Statistics South Africa (1998) Figure 5 Age profile of African migrants originating in the Eastern Cape, % Proportion 35% 30% 25% 20% 15% 10% 5% 0% African in the EC from the EC to the W C from the EC to the rest of SA within the EC Age Grouping (years) Source Census % Sample, Statistics South Africa (2004)

17 Magnitudes, Characteristics and Activities of EC Migrants 19 Considering migration into or within the provinces, Figure 6 suggests that the majority of migrants fall into the year age band in both five-year periods, with Gauteng and the Western Cape showing the highest proportions. The second most mobile age range over the two periods is the 5 19 year age band, probably as a result of children following parents who have moved earlier or possibly, for those over 16, migrating in seek of work. Those aged 60 and above were the least mobile. More generally, in the period migrants who have moved into or within Gauteng make up almost 21 per cent of the population. This is a decline from the five years prior to 1996, where they contributed 33 per cent. In the latter period, migrants who have settled in the Western Cape make up 18 per cent of the population. This also represents a decline from the period, when they made up 30 per cent. The residents of Gauteng, the Western Cape and Free State appear to be much more settled in 2001 than they were in 1996, with the number of migrants decreasing by at least 12 per cent of the total in absolute terms. However, perhaps the key point to note is that, in nearly all the provinces, the majority of the population is non-migratory. A notable exception is the Figure 6 Age distribution of migrants into or within a province as a proportion of the respective provincial population 35% 30% 25% 20% 15% 10% 60+ years years years 5-19 years 5% 0% Eas tern Cape Free State Gauteng Kw azulu- Natal Limpopo Mpumalanga Northern Cape North West Western Cape Source Census 1996 and Census 2001 Migration Community Profiles (Statistics South Africa, 1998 and 2003)

18 20 Southern African Journal of Demography 11(1) migration stream of Africans from the Eastern Cape to the Western Cape, which still accounts for a significant proportion of the expanding African population in the Western Cape (Bekker 2002). Employment and occupation Figure 7 and Table A.3 in the Appendix analyse the employment status of migrants using the expanded definition of unemployment from the census in 2001, and a consistent definition in This analysis only looks at the economically active people as defined in the censuses (i.e. between the ages of 15 and 65). Additional detail is provided in Tables A.4 and A.5 in the Appendix. The most striking observation is the general increase in unemployment rates of migrants between the period and the period. 9 More specifically, these tables show that it is likely that an African migrant out of the Eastern Cape is moving in search of work. Of the people moving into the Western Cape in the five years prior to Census 2001, 38 per cent were unemployed, while 36 per cent of those moving to the rest of South Africa fell into this category. The corresponding 1996 census data give rates of 29 per cent and 32 per cent respectively. The changes in rates between the two censuses seem to point to a worsening of economic status of those moving to the Western Cape, relative to those moving to the rest of South Africa. Figure 8 goes on to show the distribution of unemployed migrants by how soon they could start work. This figure looks at the unemployed from the 2001 census alone as this question was not asked in the 1996 Census. Almost two thirds of the African migrants leaving the Eastern Cape are willing to start work within one week. This far exceeds the proportion associated with non-migrants in, and migrants within, the Eastern Cape, where the respective percentages are 33 per cent and 45 per cent respectively. This indicates the keenness for work of those leaving the province. Three quarters of the non-migrants in the Eastern Cape fall outside of working age and, therefore, are not economically active. Only 9 per cent are employed. The higher levels of employment of African migrants both out of, and within, the Eastern Cape would seem to indicate that those with jobs are willing to undertake the risks of migrating in search of potentially better jobs. Such an explanation is in line with trends at 9 It is possible that part of the explanation lies in the slightly different questions asked in each census and the resultant regrouping of 1996 data to get responses consistent with the 2001 data.

19 Magnitudes, Characteristics and Activities of EC Migrants 21 Figure 7 Employment status of economically active migrants originating in the Eastern Cape 100% 90% 80% 70% Proportion 60% 50% 40% 30% 20% Not economically active Unemployed Employed Not Applicable 10% 0% African in the EC from the EC to the WC from the EC to the rest of SA within the EC Stream Source Census 1996 and Census % Sample, Statistics South Africa (1998 and 2004) Figure 8 Distribution of unemployed migrants by how soon they could start work 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Countrywide Migrating Non- African Countrywide Countrywide African in the EC from the EC to the WC from the EC to the rest of SA within the EC Proportion Does not choose to work More than one week W ithin one week Stream Source Census % Sample, Statistics South Africa (2004)

20 22 Southern African Journal of Demography 11(1) Figure 9 Occupational status of employed migrants, % Unspecified 90% 80% Other Proportion 70% 60% 50% 40% 30% 20% 10% 0% African Countrywide Migrating Non-African Countrywide Countrywide African in the EC Stream from the EC to the WC from the EC to the rest of SA within the EC Elementary occupations Craft and related trades workers Technicians and associate professionals Professionals Legislators, senior officials and managers Source Census % Samples, Statistics South Africa (1998) Figure 10 Occupational status of employed migrants, % Unspecified 90% 80% Other Proportion 70% 60% 50% 40% 30% 20% 10% Elementary occupations Craft and related trades workers Technicians and associate professionals Professionals 0% African Countrywide Migrating Non-African Countrywide Countrywide African in the EC Stream from the EC to the WC from the EC to the rest of SA within the EC Legislators, senior officials and managers Source Census % Samples, Statistics South Africa (2004)

21 Magnitudes, Characteristics and Activities of EC Migrants 23 the national level as shown in Figure 7 and Table A.3, where migrating Africans during both five-year periods are generally more likely to be economically active than their non-migrating counterparts, and seem to be moving either in search of employment, or in search of better jobs. Interestingly, Table 8 shows that, in 1996, Gauteng seemed to be the most popular province for people residing in the Eastern Cape but working in another province (with 35 per cent working there), followed by the Western Cape (19 per cent) and KwaZulu-Natal (17 per cent). However, by 2001, the order changed, with the majority of those Eastern Cape residents working in another province being in KwaZulu-Natal (31 per cent), followed by Gauteng (21 per cent) and the Western Cape (19 per cent). It has to be noted, though, that the overwhelming majority of the employed population of the Eastern Cape work in that province (almost 95 per cent). Table 8 Distribution of the Eastern Cape population, working elsewhere, by the province where they work Province Number Percentage Number Percentage Western Cape Northern Cape Free State KwaZulu-Natal North West Gauteng Mpumalanga Limpopo Total Source Census 1996 and % Samples, Statistics South Africa (1998 and 2004) Figures 9 and 10 show that migrants in both five-year periods are employed in similar proportions in the same occupations. Interestingly, the majority of the working African migrants from the Eastern Cape to the Western Cape (54 per cent in the period and 52 per cent in the period) are employed in elementary occupations such as street vendors, domestic workers, building caretakers, agricultural and fishery labourers, and construction, manufacturing and transport labourers. This far surpasses

22 24 Southern African Journal of Demography 11(1) any other migration stream; where typically between 33 per cent and 39 per cent are employed in these occupations. Imbalances in skills between the population groups are also apparent. For each five-year period, respectively, 23 per cent and 31 per cent of non-african migrants are legislators, senior officials, managers, or professionals, while only 8 per cent of African migrants fall into these two groups. Notable changes in proportions occurred mainly amongst the Eastern Cape population. The proportion of African migrants within the Eastern Cape, who were professionals, decreased from 12 per cent in the five years prior to Census 1996 to 7 per cent in the five years prior to Census The corresponding figures for the African non-migrating population of that province showed a similar fall from 13 per cent to 5 per cent. These decreases were accompanied by increases in the proportions of technicians and associate professionals. For this occupation group, the proportion of Africans migrating within the Eastern Cape increased from 4 per cent to 11 per cent, while the proportion of non-migrants increased from 4 per cent to 13 per cent. Of those moving out of the Eastern Cape, large changes were noted amongst craft and related trades workers. The proportion of Africans involved here, moving to the Western Cape, fell from 15 per cent to 10 per cent, while the proportions moving to the rest of South Africa fell from 21 per cent to 14 per cent. Figure 11 examines whether those migrants who are employed are employees, employers, self-employed or family workers. The y-axis has been truncated to aid with the scale of the figure with the full results being presented in Tables A.6 and A.7 in the Appendix. From this it can be seen that at least 84 per cent of any group of working migrants are paid employees. Moreover, these proportions have increased on the whole between the censuses, and are accompanied by decreases in the proportion who are employers. Also, the proportions of the self-employed have largely fallen between censuses, and are more variable in 2001 than 1996, except for migrating non-africans (increased from 7 per cent to 13 per cent). This corresponds with a large decrease in the proportion of employers in this group (7 per cent to 2 per cent), which probably reflects the change in employment status of non-africans due to transformation of the work force. The very high reliance on others as providers of work amongst the African migrants from the Eastern Cape to the Western Cape is evident.

23 Magnitudes, Characteristics and Activities of EC Migrants 25 Figure 11 Employment status of employed migrants (truncated y-axis) 30% Proportion 20% 10% Employee Employer Self-employed Fam ily worker 0% Nonmigrating African Countryw ide Migrating Non-African Countryw ide Migrating African Countryw ide Nonmigrating African in the EC Stream Migrating African from the EC to the WC Migrating African from the EC to the rest of SA Migrating African w ithin the EC 40% 35% 30% 25% 20% 15% 10% 5% 0% No schooling Some primary Complete primary Some secondary Grade 12 / Std 10 Higher Source Census 1996 and % Samples, Statistics South Africa (1998 and 2004) Figure 12 Educational classification of migrants originating in the Eastern Cape, Proportion African in the EC from the EC to the W C from the EC to the rest of SA within the EC Source Census % Sample, Statistics South Africa (1998)

24 26 Southern African Journal of Demography 11(1) Education Figures 12 and 13 present the distributions of the educational classification for various types of Eastern Cape migrants This presentation excludes all children under age five. These distributions are seen to have remained largely the same over the two five-year periods. Both figures show that the largest proportion of migrants out of the Eastern Cape has some secondary education, with the second largest having some primary education. These proportions have remained stable. However, there are some changes in the proportions of the other groups. The third largest category in the period was No Schooling, while this was Grade 12/Std 10 in the period. Overall there seems to be a slight improvement in educational attainment amongst migrants. In contrast to the relative stability of the distribution of the educational classification of the migrants, the corresponding distribution of nonmigrating Africans has changed markedly. In the 1996 census it was skewed to the left. In the 2001 census, the proportion of those with no schooling fell relative to 1996, while the some primary and some secondary Figure 13 Educational classification of migrants originating in the Eastern Cape, Proportion 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% No schooling Some primary Complete primary Some secondary Grade 12 / Std 10 Higher African in the EC from the EC to the W C from the EC to the rest of SA within the EC Source Census % Sample, Statistics South Africa (2004)

25 Magnitudes, Characteristics and Activities of EC Migrants 27 categories increased. Most simply, this could point to an increase in education amongst non-migrants in the Eastern Cape. Alternately, it could relate to the selective effect of migration, in that many people with no schooling have, in fact, left the province. This seems unlikely given that, overall, it appears that migrating Africans of Eastern Cape origin have a higher level of education than non-migrating Africans of this province. These findings seem broadly in line with the national results shown in Figure 14. Here, too, non-migrating Africans are, on the whole, less educated than migrating Africans. With 47 per cent of migrants having some secondary education or higher in 1996 and 60 per cent in 2001, it also seems that African migrants out of the Eastern Cape to the Western Cape are slightly more educated than African migrants from the Eastern Cape to the rest of South Africa where the respective percentages are 44 per cent and 58 per cent. Of particular concern is that, over both five-year periods, at least 60 per cent of the non-migrating African population of the Eastern Cape has no schooling, or only some primary schooling. This highlights the low level of education amongst the population of this province. Figure 14 Educational classification of migrants countrywide 100% 90% Proportion 80% 70% 60% 50% 40% 30% 20% 10% Higher Grade 12 / Std 10 Some secondary Com plete prim ary Some primary No schooling 0% African Countryw ide Migrating Non- African Countryw ide Stream Countryw ide Source Census 1996 and % Samples, Statistics South Africa (1998 and 2004)

26 28 Southern African Journal of Demography 11(1) Tables A.8 and A.9 in the Appendix present some detail on the qualifications of migrants and non-migrants with post-school qualifications. Such a comparison provides an opportunity to see where highly educated migrants and non-migrants perceive their skills could best be put to use. However, it should be noted that direct comparison is difficult because there were changes in the choices available to respondents in the relevant questions of the 1996 and 2001 censuses. The major share (40 per cent) of Africans in the Eastern Cape who did not migrate, or who migrated within the province, in the period prior to Census 2001, were involved in education, training and development. This proportion was even greater in the period prior to Census 1996 (52 per cent). Also, of those choosing to move out of the Eastern Cape in the period prior to Census 2001, 21 per cent were involved in business, commerce or the management sciences a greater proportion than those choosing to stay behind. Of those who chose to move (out of, or within the province) in the period , there was a greater proportion involved in engineering and engineering technologies than those who never moved at all. For African migrants to the Western Cape, it is interesting to note that the proportion involved in the computer science and data processing, or computing fields, was greater than other flows in both five-year periods. Similar comments apply to the portions of this stream involved in technical, and administrative and clerical work, in the period prior to Census CONCLUSIONS Using an empirical picture drawn from South Africa s two most recent national population censuses, this paper has provided much empirical detail on South African internal migration patterns and recent changes to these patterns. This is not a policy exercise per se in that it does not rigorously assess the causes of migration or the impact of a specific policy on migration flows. Nonetheless the findings are relevant to policy. The budgeting of and planning for service delivery and social welfare uses information on the present distribution of the population and likely changes to this distribution. This final section of paper pulls out some key findings that seem particularly germane in this regard. Countrywide, migration patterns reveal strong trends of urbanisation within the South African population. Africans in the Eastern Cape are still largely rural-dwellers. However, this group is urbanising at a faster rate than