 This is a descriptive study. It's a mixed method study. It draws on a national panel survey and also on qualitative research. And I'm really trying to just get into how our family is fragmented in South Africa. It's critical because part of the apartheid legacy was that family fragmentation was structural. So what has happened post-apartheid? Am I right? So what I'm doing here is focusing on children as a substantial but very neglected part of the migrant population. And when I say migrant population, we can see in what ways they're affected. Because they're not necessarily migrants, but they are affected by migrant labour. And there's been very little work that tries to look at patterns of child migration in the context of labour migration. They've become quite invisible participants in migration processes. So really what I'm going to do is just a few thoughts on how might migration theory and what we know as the mechanisms of migration, particularly urban migration, be considered from the perspective of children. Look at what are the patterns of child migration and do they differ from those of adults? How is child migration related to maternal migration? And I'm focusing specifically on the relationship between children and mothers, although one could do it for fathers as well. It would be more difficult because fathers are more absent from the lives of children. And then what can qualitative research add to our understanding of child migration as a part of a household strategy? And we'll see how much of that we get through. So departure points here are this long history of the disruption of family life in South Africa. And South Africa isn't anomaly in this way, but just as Colin Murray wrote from Lesotho about the effect of migrant labour on families, and using that as a kind of a proxy for also for internal migration in South Africa. So similarly a lot of this could be applicable to Lesotho, Swaziland, Zimbabwe, Namibia, to a certain extent Zambia, and Mozambique and other migrant sending countries around South Africa. The important thing is that historically cities tried to keep people who were not economically useful out. They became surplus populations who were relegated to what were established as independent homelands, which were actually literally made into separate countries. And people were forcibly removed there, and that was primarily women, children, and the elderly. The repeal of these legal constraints only from the mid-1980s meant that there was a sudden expectation that you would see massive permanent family migration. But studies that have been done have suggested this has not happened. So what are the barriers to that? We've seen that mines have declined as major employers, and those were the major migrant employers that set up hostels with single-sex hostels that weren't conducive to family living arrangements. There's been in place of this the rise of more insecure and poorly paid work in the informal sector and also in domestic work. And so alongside this we've seen a rising share of female migration as a share of labour migrants in South Africa. And this is important for children because these are prime-age women who are also women of child-bearing age. At the same time, rates of marriage already very low in South Africa and union formation have declined further. So we have women bearing a financial burden and also a care burden. And what I think is interesting is how women deal with the trade-offs and the decisions they have to make. There's been research that has shown that being a mother has a preventative effect on migration, but also conversely that labour migration is a key reason for maternal absence from children's lives. There's also been work that has shown that the presence of family members who can care for children reduces that preventative effect on women. And so when there's a substitute carer in our context, particularly grandmothers, women are more free to be able to migrate for labour purposes. There's also a connection between cash transfers and the ability of household members to migrate. And cash transfers have been expanding. And as I said, it's women aged 15 to 20, young women who are most likely to migrate. So this, of course, would affect children. And within that category, there are these categories of women moving alone, possibly leaving behind children, women moving with children, moving with men and children. But of course, children involved in all of these scenarios. So then if we look at the migration literature and start thinking about, well, in terms of what we know about the mechanisms of migration, what might it mean for children? It's established that when you have perpetuated systems of temporary or circular migration, they're reliant on family members at the home of origin to sustain ties between urban and rural nodes, to enable people to migrate and yet have the security of home. And it's been suggested, in fact, that the sort of spatial dispersion is not necessarily a fragmentation of the family, but a strategy to conserve a family. The whole principle of cumulative causation that people tend to migrate from one area to the same area because of social networks or family members, places that can provide initial accommodation. But that could work against children because where one is dependent on the generosity of other people in social networks and where places are overcrowded, one can't necessarily bring dependence with one. Informality is a mechanism, as a stepping stone to the city. And so people often migrate to the periphery of cities into informal environments. But again, that's a risk that perhaps labour migrants are willing to take, but may not be prepared to take for their children and there's a lot of work on the dangers of informality, the absence of social infrastructure, the risk of fire. You'll see stories coming out about that. Processes of chain migration, which is partly what you're referring to, where you have sequential migration of family members. In the international literature, there's been some work that looks at different patterns, different sort of typologies of chain migration when it comes to children. So you might have co-migration where they migrate simultaneously, or sequential migration where it's delayed, but they follow. You might have reverse migration where children are sent away from the home of a labour migrant. You might have non-migration where children are left behind. You might have autonomous migration where somebody migrates and the child migrates somewhere else. You also, in relation to that, this is, I think, an interesting idea of involuntary immobility. So the idea of lacking the capacity to be able to migrate. And we see this quite a lot at Camus-Arton and South Africa, where there's an intention of sequential migration. But there are barriers to that. And so from the perspective of children, you can have a long period of limbo, which might last an entire childhood in which the aspired to family arrangements are never realised. There's also, in the work of anthropology, been really nice work that broadens from the kind of individual agency rational choice models to consider household strategies as a meso-level of focus between kind of the structural issues and the individual level issues. And a child-focused perspective helps with this because it helps us to see beyond the household. And those questions of when households are dispersed, how do we understand relationships between them. So firstly I'll start with a little work that is in panel data. We've got quite a data-rich country. We have censuses, cross-sectional surveys, an annual general household survey, but they're limited use for migration analysis because they rely on recall. And I'll get back to that. Also migration questions have been prioritised in these surveys, so we just don't have the kind of detail on migration that we might want. We've had two demographic surveillance sites which have ongoing studies with panels, and a lot of really good migration workers come out of that. So we're starting to see interesting dynamics come from them, but a major problem, of course, if one's thinking of internal migration across the country, which is often across province, is that once people move out of the site, they lost the panel. So, hurrah! We have a new national income dynamics study, which is our first nationally representative panel survey. And there were four waves of data available, so the baseline in 2008, then again 2010, 2012, and 2014-15. And I've used these four waves. At baseline, it's a nationally representative sample of 7,300 households, nearly 30,000 individuals, of whom 9,600 children under 15 have been restricted to what are defined in the surveys as African children under 16. We still have these categories in our surveys, and it's partly because in this instance I'm interested in children in the context of historic legislation that was discriminatory. So if you put everyone in the mix, you will start seeing confounding patterns because wealthy people migrate for very different reasons, and people who have been historically constrained to rural areas would migrate for different reasons. I then had to restrict it to only children who were under 8 years in wave, one of the data so they could still be under 15 in wave 4, and I could still have access to all the information in the wave 4 child module. And there's some attrition, and it's a work on attrition, but it doesn't really affect the variables that we're interested in. It left me a balanced sample of 3,750 children to work with. Migration can be defined in many ways. I define it as a move across districts, so a physical move across districts, and it's not relying on reported migration. I can actually see which district children live in, and I can look for changes in districts. And according to that definition, 14% of this cohort of children had geographically migrated over this 7-year period. I then also was able to map mothers onto children to explore maternal and child migration events and co-residents. And migration rates for mothers of children was part of the sample that you couldn't tell it for because they were absent in all waves and they might have migrated or they might not have, but you can't tell. But that rate was around 17% if you excluded those. So fairly high rates of migration for both children and mothers bearing in mind just a cohort over a small period of time, and this is geographic migration. Another benefit of NIDS, the National Income Dynamics Study, is that it uses a broad definition for household. So what household is depends on how it's defined in the household roster, right? So, commonly, it has a kind of a subjective aspect, which is people who consider it to be part of the household. And then you can add on to that a narrow definition, which is what happens in most of the official surveys of people who stay together for at least four nights a week. And whereas a broad definition includes household members, people who've stayed here at least 15 days in the last year. So that is allowing for labour migrants after the household. And then the third aspect is the whole idea of a common resource pool or eating from the same pot or whatever. So in South Africa, certainly, the broad definition is how labour migrants are defined, i.e. non-resident members at a household of origin, because it is meant to reflect an attachment to the household of origin and some pattern of return. Now, I mentioned the other countries around South Africa. They're incredibly high rates of parental absence from children's lives. And we use 1993. This was a previous Living Standards Survey, a national survey that was conducted in South Africa. And that can be taken as a kind of a baseline at the end of apartheid. What has happened since, what we see is a decline in the share of children living with both their parents, an increase in the share of children living with neither parent. I think only Zimbabwea has higher rates of parental absence from children's lives out of a study of 77 countries in the world. And the other thing we see is quite a high rate of maternal involvement in... well, high rate of co-residents between mothers and children, often in the absence of fathers. So what has happened since apartheid is not greater patterns of co-residents of children with their parents, but in fact less so. Sorry. When I look at, and I'm trying here to distinguish between the difference between non-residents who we could take as being labour migrants and absent parents. And the lovely thing about working with children is even when somebody is absent from a household and would never appear on the household roster, you've got some demographic information about the parents of children who are absent. And here, just trying to give a sense of when parents are absent, are they actually absent from the lives of children? No, generally not. We can see that in fact very few children with absent mothers never see their mothers. Okay? A slightly higher proportion of children with absent fathers or quite 30% or 31% of children with absent fathers don't see their fathers. They may be children whose fathers are not even on their birth certificate. We see a slight difference between non-residents, in other words, labour migrants, in both the case of mothers and fathers. And we see more financial support to children from parents who are non-resident household members or as a proxy for labour migrants. So, the next thing we do is we look at where children are moving in the country. So, this is once we've defined our migrant children. Oh, gosh, okay. I'm comparing this to work that has been done by other people at UCT who looked at adults and what we saw was that the modal direction of movement is actually within area type. So, people are moving to a different city or a different rural area, but it's largely within those. And what we see for children, and this is from the child panel analysis, is similarly a kind of a modal direction in the same. But what I want to draw your attention to is the very difference. So, we're seeing a much more kind of cross area type movement for children. Children's migration patterns differ from those of adults. We also, just as a simple logic question which shows that the migration of children is strongly related to the migration of mothers. Okay, so children, if mothers migrate, children are more likely to migrate because it's also some relation to work-seeking activity of mothers, but they don't necessarily migrate at the same time or in the same direction. So, what it then does is take children and mothers together and create a kind of a typology of migration events where either the child moves or the mother moves or they both move. Okay, and here we can start getting into those sort of familiar categories of children left behind and so on. And what we see is that nearly half of the migration events result in children living together with mothers, but more than half of the events result in children living separately from mothers. So you've got your children sent away, children left behind, autonomous child migration as opposed to, and the biggest, the other biggest segment is where you have this sort of chain or sequential migration of children to join their mother. This simply tracks, let me just see, the kind of the directional, urban-rural direction of these migration events. So where mothers and children stay together or stay apart, there's actually very little difference between they could be going, the direction could be in the direction of urban or rural. But when they are moving apart, the direction is primarily of the child, is primarily to rural areas. And when they're moving to unite or when they end up together, the direction of child movement is primarily to urban areas. So what we're seeing is that rural households continue to carry a huge burden of support for the children of migrant workers, which was the case historically, when migrants are men, it's now the case when migrants are women. I don't think I have time to get into the case study. You can read, there's a foreshortened version in the... I have two minutes. Well, the case study, it took nearly two years, just looking at... This is the complete opposite, right? One migrant mother and her extended family spanning three generations. And from a particular township in the western Cape, which I could see through a migration cell that was conducted there, is linked through these cumulative causation networks, these social networks to a very specific village in the former trans sky, which is part of the independent homeland. Historically, you would literally have to go through a border post to get from the trans sky to Cape Town, even though this is all one country. It's very rural, it's a scattered village. Her name is Lindy Ware, you'll see it here. So, I mean, using the 100% census, I could get down to this tiny village of only 1,200 people. And what you see is a huge migration of adults from that village. If you compare that to the urban township where she lives, it's a completely different picture. Okay, so you have... You're seeing children born, you're seeing an under-representation of children, you're seeing a large young migrant population, and then absolute tailing off as elderly people return, or it's also a fairly new settlement. So, what I wanted to talk about was partly about the method and the usefulness of doing this kind of research. Problems of recall happen in qualitative research, too. The initial story I got was turned out to... I mean, it changed every time we spoke. It was contradictory. There were two main methods we used to assist recall. One was literally a family history grid with columns for each person, rows for each year, and so we'd start aligning them and going, and we'd say, okay, we're going to have a marriage. Hang on. If that happened, then this person must have been living in... We're looking at who's sharing households. So we could keep on working with this thing, and as we did, her memory was refreshed. We also worked a lot with kinship diagrams. So here's a diagram of her childhood household. Here she is, the second wife of her husband, her half-brother from her previous father, her other siblings, grandchildren from a former marriage. Okay, so this is what we would call a complex extended three-generation household. I'm gonna stop here. But for her, it wasn't complex. It was quite normal in her environment. And then what we start doing is tracking people moving out. Wangani, first, the father is murdered over a stolen horse. Lin Diwe is raped and gives birth to her first child at the age of 15. The oldest son is sent to Cape Town to go and earn money, but it's not enough. Her mother starts migrating in and out of the household between Cape Town and back. She gets too tired to do that, and the son gets sick in Cape Town. And so the mother then sends Lin Diwe, who by now has given birth to her second child, to Cape Town to go and help support the family. So this is not an individual migration choice. This is an absolute household strategy. Sorry, we walk around the village quite a lot. I can't get into how her childhood strengthened her sense of rootedness to the place. But we have these really interesting notions of how there's an absence of choice. You stand up and you simply have to act to remedy a situation. You have to leave your children behind. There isn't an option. This is her mother talking about leaving a child-headed household behind. Lin Diwe similarly talks about having to, being forced to leave the house in order to go and earn income and prepare for those who come after her. And so I won't go through this. You'll just have to believe me when I say the move to an informal settlement in Cape Town, multiple shack fires, delays in the children joining, but actually the children moving there before she's found the house that she'd wanted to live in. So they're exposed to huge risks having to share with families and then family tensions and moving out and establishing new shacks. Four shack fires while they're living with the children in which they separate into different households. Finally, acquiring a formal house in the township, her mother's return to the eastern Cape. We just see an absolute, this incredible fluidity of movements and some intention to return to the rural area. The conclusions then that I will draw which mostly come out of the case study, which you haven't heard. But what we're seeing is the necessity of female migration even at the cost of fragmenting the family. We see the child migration maybe prevented. So you have involuntary immobility or it may be delayed, or in fact in this case it was premature. The housing situation was still risky but the children needed to go to a better school to get a better education than one she'd had in the rural areas. So people are constantly having to revise their plans and aspirations. But both migration and immobility, non-migration may be about a lack of choice. And this challenges notions of individual agency which so often informs studies of the causes of migration. But we see that long-term intentions are superseded by short-term necessity. We see the importance of extended families, the absence of men, the role of grandmothers, the connectedness of rural and urban homes. Another thing that emerges is the distinction between permanent and temporary migration, which we never managed to establish in the case of this one woman because intentions vary all the time depending on experience. The surveys are essential for describing broad trends but they're not well suited to examining extended families, dispersed families, household arrangements or social networks. And the child and mother migration events take many forms and are worthy of more research. Thank you.