WPS6200 Policy Research Working Paper 6200 Migration and the Transition to Adulthood in Contemporary Malawi Kathleen Beegle Michelle Poulin The World Bank Development Research Group Poverty and Inequality Team September 2012 Policy Research Working Paper 6200 Abstract In many African countries, the timing of important life to explore socioeconomic and demographic aspects events—such as school-leaving, first marriage, and entry of youth transitions to adulthood and which tracked into the labor market—is thought to be strongly tied respondents as they moved to new dwellings. Among the to migration. This paper investigates the relationship sample, they find that moves are not uncommon, and between major life events, household characteristics, the predominant reasons for moves are non-economic. and migration among adolescents and young adults in Although historically ethnic traditions in this area have contemporary Malawi. The specific research questions held that girls and women usually did not move upon are twofold. First, what are the socio-economic and marrying, the data show that women were more likely demographic determinants of migration? Second, how to move between survey rounds than boys and men, and do school attendance, first marriage, and employment- that marriage was the main reason for doing so. Closer seeking relate to migration patterns? The study uses ties to the head of the household are associated with less panel data collected from a survey designed specifically movement for both women and men. This paper is a product of the Poverty and Inequality Team, Development Research Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The author may be contacted at kbeegle@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Migration and the Transition to Adulthood in Contemporary Malawi Kathleen Beegle Michelle Poulin JEL codes: O15, O12 Keywords: internal migration, youth, gender, Malawi Sector board: Poverty Reduction Kathleen Beegle is at the Development Research Group, The World Bank and Michelle Poulin is at Stanford University. The views expressed here do not necessarily reflect those of the World Bank or its member countries. The authors are grateful for comments from at the IUSSP Seminar on Youth Migration and Transitions to Adulthood in Developing Countries and from an anonymous reviewer. I. Introduction The set of decisions made in the transition to adulthood -- including when to exit school, when to marry, and when to seek employment -- are critical with respect to a young person’s well-being over the life course. These life events are, in turn, related to moves to new dwellings and new communities. Consequently, migration is common in adolescence and young adulthood, often more so than for older adults. Yet empirical studies of migration tend to focus on adults, and not specifically adolescents and young adults. It is a topic sometimes omitted from major studies on transitions to adulthood or only touched upon in a limited fashion often due to data limitations, such as the Panel on Transitions to Adulthood in Developing Countries (2005). Equally important, work on migration tends to emphasize international migration which for many developing countries captures only a small fraction of migration. Migration for many is not only within-country, but may be fairly local. The objective of this paper is to investigate the relationship between household characteristics, major life-cycle decisions, and migration among adolescents and young adults in contemporary Malawi. Our research questions are twofold. First, what are the socio-economic and demographic determinants of migration, and how are these conditioned by gender? Second, how does school attendance, first marriage, and employment-seeking relate to migration patterns? In a departure from many migration studies in Sub-Saharan Africa, we emphasize moves that are the most common and therefore the most relevant: within-country and often local moves to neighboring communities. To investigate these questions, we draw on two data sources. We examine descriptive patterns from cross-sectional national household survey data. For further analysis, we take advantage of panel data collected from a survey designed specifically to explore socioeconomic 2 and demographic aspects of youth transitions to adulthood, which enable us to examine the impact of pre-migration household characteristics on moving. We use these unique data to compare the characteristics of those who migrate with those who do not, including how these two groups differ on the timing of three key life events--leaving school, entering the labor market, and marrying for the first time. We examine whether and how these patterns differ for young males and females. . These analyses are part of a broader agenda to understand aspects of transitions from youth to adulthood in a low-income setting. II. Migration in Malawi Researchers of migration often categorize migration into distinct types such as international, internal, temporary, seasonal, and circular (versus linear), among others (see discussion in Pastore 2008). Population mobility, especially for the working-age, is a critical feature of traditional economic models of development, including the Lewis model and the Harris-Todaro model (Lewis 1954; Harris and Todaro 1970). The ‘new economics of migration’ (Stark and Bloom 1985) emphasize how migration is part of a general livelihood strategy for the initial household as a whole. In this model, movers are part of a welfare maximizing strategy for overall household income growth, as well as for risk sharing and easing credit constraints in the presence of incomplete or missing markets. For example, Rosenzweig and Stark (1989) find that migration patterns for marriage in rural India are consistent with risk-sharing strategies of the initial household. In this scenario, migration has benefits for those who do not migrate. Even if non-movers do experience benefits from another’s migration, however, evidence from Tanzania shows that migrants nonetheless do significantly better economically as a consequence of their move, compared to their non-moving counterparts (Beegle, De Weerdt, and Dercon 2010). 3 Yet, even while migrants fare better economically, people migrate for reasons that may not be clearly linked with economic gains, such as moves associated with marriage. In our country of study, Malawi1, the largest ethnic group is the Chewa, who historically practice matrilocal patterns of social organization—husbands have typically moved to their wives’ compounds following marriage, which is often in other villages. Over the past century, however, matrilocality has become more irregular. The extent to which the Chewa continue to practice matrilocal forms of relocation has been on the wane in recent decades, with the increasing role of labor migration cited as a potential reason (Phiri 1983). We will return to this topic in our empirical analysis below. Although studies of migration and remittances in Sub-Saharan Africa are numerous, few have focused specifically on the population of young adults. One exception is McKenzie (2007), whose research demonstrates that a large share of migrants is constituted by young people; he concludes that their reasons for and experiences related to migration differ from migrants who are slightly older or much older. However, McKenzie’s work considers international migration only, which currently, at least in Malawi, is a relatively rare form of migration compared to internal moves. Sander and Maimbo (2003) characterize migration in Sub-Saharan Africa as unique, since most people on the continent move intra-regionally and within-country, rather than overseas. Although during Malawi’s colonial era international migration was common, following independence in 1964, President Kamuzu Banda began to discourage international migration, 1 Malawi is an impoverished small country in southern Africa. Its population of 13 million is 85% rural (NSO, 2008), with most people living from subsistence farming supplemented by small-scale income-generating opportunities that are typically more available to men than they are to women. Agriculture accounts for 34% of its GDP, largely generated from smallholder farmers, with its main products being tobacco, tea, coffee, and cotton. The country is poor even by African standards: in 2007 GNI per capita was $250 compare to an average of $952 for the Sub-Saharan region (World Bank 2010). 4 particularly to South Africa, to keep workers in the country . He formalized this sentiment in 1974 with economic policies designed to build the country’s cash-crop estate agriculture. Those policies proved effective: data show that in early 1972, 10 percent of the population resided abroad (Boeder 1974), but a significant portion returned in following years (Christiansen and Kydd 1982), so that by 1975 labor migration to the mines in South Africa had nearly ceased; the number of Malawians under contract with South African mine labor organizations declined from 123,845 to 2,711 during this period (Ministry of Labor 1977). In the early 1980s, as export prices declined and drove down demand for migrant labor in South Africa, migrant labor also fell (World Bank, 2010). Migratory movement to South Africa did expand again in the 1990s, although not under formal contracts (Andersson 2006). In more recent years, Malawi has witnessed a massive increase in internal migration, both rural-to-rural and rural-to-urban, and the bulk of migration experiences for Malawians today remain internal (Englund 2002). Mtika (2007) identifies circular migration as a main economic survival strategy for prime-age Malawians, although he does not clearly define it or quantify the extent to which it occurs. Aside from the economic motivations for migration or its consequences, there are other implications of migration for people’s well-being. Anglewicz and Reniers (2010), for example, study the links between divorce and migration among adults in Malawi, and find that divorced women are more likely to migrate. They also find that migrant women are more likely to get divorced. Migration is also thought to be closely linked with health. Internal labor migration has been identified as a contributing factor in the spread of HIV/AIDS, both in Malawi and indeed throughout southern Africa . IOM (2008) outlines several of the relevant industrial sectors associated with migrant workers and the implications for the HIV/AIDS epidemic in Malawi. 5 Migration in Malawi is common. Data from the most recent nationally representative survey in the country (the 2004/05 Integrated Household Survey, IHS2) show that slightly over half of adults (15 and older) have not always resided in their current location. About one in five of adults moved into their current village or neighborhood in the last five years. Among working- age adults, migration rates are even higher (see Table 1). Among those who moved into their current community in the last five years, the vast majority (more than 90 percent) moved from within the same district. So while moving is common, these are not long distance moves. The focus in this paper is on young adults, but of note is that migration is also observed among children. Child fostering is not uncommon in Sub-Saharan Africa and it implicitly results from migration of either parents or children. Malawi is no exception. Ansell and van Blerk (2004) suggest that internal child migration in Malawi has historically been a coping strategy for families dealing with stress and that its importance has increased due to the HIV/AIDS epidemic. While orphanhood rates have increased, even among non-orphans in Malawi there has been an increase in propensity of non-orphans to reside without either parent (Beegle et al. 2010). III. Young Malawians on the Move In this section we focus on the descriptive evidence on migration among young adults in Malawi, a population facing complex life-cycle events. We do not present a comprehensive framework or model on the transition to adulthood in Malawi. Rather, we hope this analysis serves as a starting point for such a framework, noting that the main features of the transition to adulthood center on marriage, work, and school, all of which are linked to local mobility. In the case of secondary boarding school, this mobility may even be quite distant. 6 At the time of the IHS2 survey (2004/05), 43 percent of young people 15-24 years of age had resided in another village or neighborhood (Panel A, Table 1).2These numbers show how common geographic mobility is among young people in Malawi. Compared to working-age adults, rates among adults over 50 years of age are lower. About one-fifth of working-age adults have resided in their current communities for fewer than five years (Panel B, Table 1). Among younger Malawians (between 15 and 24 years of age), the rates are even higher, showing that younger adults move with high frequency. Nearly one-third (33%) of women aged 15-24 have been in their current community for fewer than five years, a finding consistent with marriage- associated migration; for men this statistic is 23%. Many of these moves occur across district lines, as shown in Panel C of Table 1. Among men in particular, almost half of movers had resided previously in another district. Never-married women and men are less likely to have resided in a different community or village, compared to their ever-married counterparts. Among never-married men, 80% resided in the community for five or more years, compared to 66% of married men. For women, the rates are 82% and 57%, respectively. This suggests that marriage contributes to relocation. Ever- married young movers are more likely to have moved within the district (more locally), whereas among the never-married, a larger share of moves are across districts. Young adults report a variety of reasons for their moves. Generally, local moves are dominated by marriage for men and women. These moves are within the district and therefore 2 The examination of rates of migration across age groups captures two effects: aging and cohort effects. The former captures the effect that older people have had more years exposed to the possibility of a move; once a person has moved, s/he is forever identified as having a migration experience. Cohort effects are the idea that processes and patterns than may evolve over time and not be driven by age but be associated with one’s cohorts.. In the case of Panel A in Table 1, the lower rate of mobility of women in the oldest age group (in terms of those who have not always resided in the current location) compared to prime-age women is evidence that propensity of a woman to move out of place of her birth has increased over time in Malawi. 7 typically shorter. Reasons for migration vary by distance of moves, especially for men, in particular when comparing work and family-related migration. Among men who moved in the last five years (regardless of marital status), moves within the district are less likely to be related to work than longer moves. About a quarter of those who moved across district boundaries reported moving for work-related reasons, compared to only 14% of men who moved within district. Work-related moves are more common among men than among women where rates peak at 7% for cross-district moves. Men are less likely to move across districts for a spouse than women (4% compared to 49%). But an additional 40% of male movers moving from a different district reported moving to live with a relative, compared to 23% of women moving from a different district. It is possible that these movers are also attending school or working in their new location. Among the subset of never-married young movers, slightly less than half of men and women moved to live with a relative, irrespective of whether the move occurred within or outside of the district. Across all groups, other reasons, such as seeking land, moving for school, returning from work elsewhere, divorce, family quarrel, illness recovery, and death of relative were very infrequently reported by young adults. Because the IHS2 was not focused specifically on migration, and because the survey is cross-sectional, these data limit our understanding of how much migration is linked to childhood (moving with parents), schooling (since secondary school facilities are not in most villages), marrying (moving to spouse’s home village), or is economically motivated (such as job search). With the IHS2, we cannot explore how geographic mobility is correlated with changes in other circumstances, such as dating and sexual behaviors, or employment, since we do not observe these data before and after the move. 8 Between 2007 and 2009, the Marriage Transitions in Malawi (MTM) project collected innovative, longitudinal data from a random sample of nearly 1,200 initially never-married women and men in the Central Region of Malawi, ranging in age from 14-20 for females and 17- 25 for males. The study was designed to understand the links between pre-marital relationships and sexual activity, the transition into marriage, socioeconomic status, and HIV/AIDS. Respondents provided detailed information on socio-economic characteristics, marriage and fertility, and sexual partnering. Two particular features of this longitudinal data set stand out. First, respondents were interviewed at short intervals, up to five times within a 24-month window. Most panel studies in Sub-Saharan Africa conduct survey rounds at a minimum of yearly intervals (and often longer), which necessitates a reliance upon retrospective reporting of events, and may bias estimates due to recall error (such as on dating and marriage).3 Second – and most relevant for this study – respondents who left the sample, due to relocation outside of the study site, were tracked by the research team. Such tracking allows us to identify the socioeconomic characteristics and demographic behaviors of mobile respondents. There are multiple ways to define migration. Using the MTM data to explore the socioeconomic dimensions of migration, we take several approaches to measuring migration. In the most general approach, we look at the distance in kilometers (based on GPS data) between the baseline (summer 2007) and follow-up (summer 2009) residence. That is, we compare the location of physical dwelling in which our sample of young adults resides using objective measures of location (rather than self-reported distances from baseline). A second approach is to look at moves out across administrative/economic boundaries, such as villages, Traditional 3 The issue of recall bias in reporting events is also relevant for migration data. For example, lacking panel data which follows movers over time, studies may need to use life history calendars to collect information on moves, the timing of moves, and the reason for the move. This is the approach of Reed, Andrzejewski, and White (2010) in their study of how migration links with education, employment, marital status, and childbearing in Ghana. 9 Authorities, and districts in the case of Malawi. We categorize moves into those within the community/village, moves out of the community but within the district, and moves out of the district. We focus on respondents’ physical moves, as defined by the distance of the move from respondents’ baseline location given any move, regardless of changes that may have occurred in household composition. Additionally, within villages, a young person may have moved to a new dwelling. By that criterion, we observe an even larger share of young adults ‘moving.’ For those young people who have not, but stayed in the same dwelling, we observe a lot of change in the composition of their household. That is, many non-movers are exposed to others who move in or out of the dwelling. For example, about one-third of non-movers have a change in household headship between the first and last round of the MTM survey. We find a considerable amount of movement over the course of 24 months. Just under a third of all respondents relocated to a village more than one kilometer away, and a fifth moved further than 10 kilometers from their baseline locations. Among those who moved 10 kilometers or more, about half of them moved across districts. Respondents report several reasons for relocating. Our tracking experience indicates that many of these moves of adolescents and young adults may not fit well with traditional economic theories of migration. For example, moves can be temporary (e.g., a visit to a relative, lasting a few months) or, for women for whom there are few labor market options, not related to employment. Still, even if moves are not made with the sole purpose of formal employment, the moves can affect labor-related outcomes, as well as other important life events that affect economic well-being, including marriage and schooling. Men were far more likely than women to move for work-related reasons (22% versus 3%, respectively). The top three reasons reported 10 for moving among women were school (7%), following parents/relative (29%), and marriage (49%). Among men, following parents and relatives were also frequently reported (35%), but compared to women, marriage was less so (21%). Using our three rounds of survey data, we can also examine sequencing of moves further from the baseline community. Table 2 shows the composition of our sample with respect to residence in Round 2 (2008) and Round 3 (2009), compared to the baseline location. Among those who moved to a nearby village from 2007 to 2008, 90% continued to reside in a nearby village. That is, 10% of these young adults moved even further away from the baseline between 2008 and 2009. Table 3 compares characteristics of migrants and non-migrants at baseline and at the end of the survey, during the final round in 2009. Several differences among the two groups are apparent. First, respondents living in a household headed by one of their parents are significantly less likely to move, compared to those who are other relatives of the head. This is this case for both men and women. Among men, 77% of non-movers were the child of the head, compared to 59% of movers. Among women, these numbers are 79% and 67%, respectively. Other relationships to the household head seem to matter for men, and not to the same degree as to women. For instance, men who are nephews or siblings of household heads are significantly more likely to move than those who were not, but the effects of being a niece or sibling are not statistically significant for women. Young men with both parents alive are more likely to be stayers compared to those with at least one parent deceased, but there is no difference in orphanhood rates by migration status for women. Among both women and men, those in the wealthier households at baseline (the highest wealth quartile) are more likely to move between baseline and follow-up, suggesting that there is 11 a cost to migration or that the wealthier have better migrant networks. Men whose main activity is farming household plots are significantly more likely to stay within their same village, whereas men who are report no economic activity was their main work but report domestic work are more likely to move. If we turn to characteristics at follow-up (the lower panel of Table 3), a striking difference is seen: female movers are twice as likely to be married by 2009. Given the predominance of the matrilocal Chewa and Yao ethnic groups in this sample (over 80%), this finding is rather surprising. In this panel, we see again the effects of being a child of household heads. While for all young adults, household size is smaller in the last round (reflecting entry into marriage and maintaining a new household), movers reside in smaller households compared to non-movers. After their move, movers continue to reside in wealthier households compared to those who did not move. Female movers are less likely to have been currently attending school at follow-up than are stayers, consistent with the finding that adolescent girls and young women who move are more likely to be married by 2009. It is extremely rare for a woman to continue to attend school once married in Malawi. To isolate the correlates of migration, we regress migration outcomes on the same set of characteristics presented in Table 3. We examine two measures of migration: a binary indicator of any move more than 1 kilometer from the baseline dwelling, and a continuous variable for the distance of the new location from the baseline, in kilometers. In Table 4, we present results from the probit and OLS estimations. Three interesting effects can be seen. First, females are statistically more likely to move at all, controlling for other observed characteristics. This is consistent with the rates of migration presented earlier. This is rather surprising—among young 12 adults, it might be expected that men are more likely to migrate for work, and especially in a traditional matrilocal community such as the Salima district of Malawi. The positive female effect in the probit model indicate that net of other important characteristics likely to be associated with migration, on average, a woman is 14 percentage points more likely to move at least one kilometer, compared to male counterparts. Controlling for household wealth does not weaken the effect. Taking our alternative definition of migration (kilometers), in the OLS model gender is not significant. This suggests that men who move, move further than women, even if women are more likely to move at all. The second significant finding is the effect of being the child of the household head, which is negatively associated with migration, an effect that is apparent in both the probit and OLS models. These findings are consistent with recent research in Tanzania (Beegle, DeWeerdt, and Dercon 2010). A few interpretations are possible, all centering on the strength of ties among household members at baseline. First, a core respondent who is not the child is (by definition) less of a dependent of the head, such as in the case of a niece, sibling, etc. Non-biological children of household heads may receive less financial and emotional support from the heads, relative to biological children of household heads. Second, the effect may be an indication that those who are not biological children are somehow more economically and socially independent from the household, therefore facilitating a greater ease of moving outside of that household. It is possible that these young people moved prior to baseline, since most children in Malawi do grow up with a biological parent as their head of household. Another possibility is those who are not children of the household heads are less likely to inherit land, and this lack of access to land 13 encourages people to move; conversely, the promise of inherited land encourages young people to stay.4 Finally, controlling for other covariates, we continue to find that household wealth is strongly correlated with moving, for both migration measures. Moreover, there appears to be some relationship with wealth and the effect of being children of household heads when moving is modeled using the distance measure in the OLS model. While suggestive that young people who lack financial resources will be constrained in their ability to move for economic or social opportunities and potentially trapped in poverty or remote rural areas, further analyses are needed to better interpret this finding. IV. Conclusions Migration studies in Sub-Saharan Africa often focus on prime-age adults and their international moves, with few studies on the patterns of mobility among adolescents and young adults. Yet, young people are often highly mobile, with moves that are often domestic and local. This study attempts to fill this gap by drawing on household survey data from Malawi. It is a first step toward developing a framework on the role of spatial mobility in the transition to adulthood for young men and women in Southern Africa. Drawing on new, panel data from Malawi, three important patterns on the migration experiences of young adults emerge. First, we find that geographical movement of adolescents and young adults is common, but young women are significantly more likely to migrate than are young men. Within a 24-month span, our data showed that just under a third of young women 4 As noted earlier, a dominant reason for moves was following parents/relatives. The lower propensity for children of the head to move shown in Table 4 suggests that these young adults move to relatives as opposed to these being family co-moves (entire groups or subgroups of households moving to new dwellings). 14 and a quarter of men moved more than one kilometer from their dwelling at baseline. Second, although men commonly reported moving for work-related reasons, the predominant reasons for migration are non-economic. Typically women move because of marriage or just after marrying, with their migration more likely to be influenced by school attendance and wealth than by age. This differs from what is expected based on the historical patterns of the dominant ethnic group, the Chewa, who traditionally practiced matrilocal residence after marriage. Finally, these data show that adolescents and young adults who live in households headed by a parent are less likely to move than are other adolescents. This suggests that stronger ties to the household head in terms of financial and emotional support are, not surprisingly, important factors influencing mobility. 15 References Andersson, Jens A. 2006. Informal moves, informal markets: International migrants and traders from Mzimba District, Malawi. African Affairs 420: 375-397. Anglewicz, Philip and Georges Reniers. 2010. Remarriage and migration in rural Malawi. Mimeo. Ansell, Nicola and Lorraine van Blerk. 2004. Children's migration as a household/family strategy: Coping with AIDS in Lesotho and Malawi. Journal of Southern African Studies 30(3): 673-690. Beegle, Kathleen, Joachim De Weerdt, and Stefan Dercon. 2011. Migration and economic mobility in Tanzania: Evidence from a tracking survey. Review of Economics and Statistics 93(3): 1010-1033. Beegle, Kathleen, Deon Filmer, Andrew Stokes, and Lucia Tiererova. 2010. Orphanhood and the living arrangements of children in Sub-Saharan Africa. World Development 38(12): 1727-1746. Boeder, R. B. 1974. Malawians Abroad: The history of labour emigration from Malawi to its neighbors, 1980 to the Present Day. PhD diss., Michigan State University. Christiansen, Robert E. and Jonathan G. Kydd. 1982. The Return of Malawian labour from South Africa and Zimbabwe. Journal of Modern African Studies 21(2): 311-26. Englund, Harri. 2002. The Village in the city, the city in the village: Migrants in Lilongwe. Journal of Southern African Studies 28: 137-154. Harris, John, and Michael Todaro. 1970. Migration, unemployment and development: A Two- sector analysis. American Economic Review 60(1): 126-142. IOM International Organization for Migration. 2008. Briefing note on HIV and labour migration in Malawi. Lewis, W. Arthur. 1954. Economic Development with Unlimited Supplies of Labor. The Manchester School 22(2):139-91. Panel on Transitions to Adulthood in Developing Countries. 2005. Growing Up Global: The Changing Transitions to Adulthood in Developing Countries. Lloyd, Cynthia B. Ed. Committee on Population and Board on Children, Youth, and Families. Division of Behavioral and Social Sciences and Education. Washington, DC: The National Academies Press McKenzie, David. 2007. A Profile of the world’s young developing country migrants. World Bank Research Working Paper Series 4021. 16 Ministry of Labor. 1977. Statistical Yearbook National Statistical Office, Malawi. Mtika, Mike Mathambo. 2007. Political economy, labor migration, and the AIDS epidemic in rural Malawi. Social Science and Medicine 64(12): 2454-2463. National Statistical Office. 2008. 2008 Population and housing census results. Zomba, Malawi Pastore, Ferruccio. 2008. Circular migration. Background note for the meeting of experts on legal migration in Rabat, 3-4 March 2008. Phiri, Kings M. 1983. Some Changes in the matrilineal family system among the Chewa of Malawi since the nineteenth century. Journal of African History 24: 257-274. Reed, Holly, Catherine Andrzejewski, and Michael White. 2010. Men’s and women’s migration in coastal Ghana: An event history analysis. Demographic Research 22(25): 771-812. Rosenzweig, Mark, and Oded Stark. 1989. Consumption smoothing, migration, and marriage: evidence from rural India. Journal of Political Economy 97(4): 905-926. Sander, Cerstin and Samuel Munzele Maimbo. 2003. Migrant labor remittances in Africa: Reducing obstacles to developmental contributions. World Bank Africa Region Working paper Series No. 64. Stark, Oded, and David Bloom. 1985. The New economics of labor migration. American Economic Review 75(2): 173-178. World Bank. 2010. Malawi country economic memorandum: Seizing opportunities for growth through regional integration and trade. Report No. 47969-MW. 17 Table 1: Residential Relocation in Malawi, by age and gender Male Female Panel A % that have not always resided in current village/community 10-14 years old 23.9 27.7 15-24 years old 37.8 47.1 25-49 years old 61.5 58.6 50+ years old 61.7 49.2 Panel B % residing in current village/community for 5 years or less: 10-14 years old 14.8 17.9 15-24 years old 22.6 32.7 25-49 years old 25.3 18.7 50+ years old 9.8 5.7 Panel C Among those who have resided 5 years or less, % who previously resided in another district: 10-14 years old 44.4 46.4 15-24 years old 49.5 38.3 25-49 years old 46.3 42.3 50+ years old 43.9 36.7 Note: Source is the IHS2 2004/05 sample of 34,181 individuals, population weighted. 18 Table 2: Movements from Baseline to Rounds 2 and 3 Location in Round 2 compared to Round 1 (baseline) Same Same Location in Round 3 dwelling community, Nearby compared to Round 1 as in but new village/ Elsewhere Elsewhere (baseline) Round 1 dwelling community in Salima in Malawi Same dwelling as in Round 1 73.7 0.0 0.0 0.0 0.0 Same village/community, but 10.9 87.0 0.0 0.0 0.0 new dwelling Nearby village/community 5.8 3.7 89.7 0.0 0.0 Elsewhere in Salima 4.8 1.9 10.3 91.3 0.0 Elsewhere in Malawi 4.9 7.4 0.0 8.8 100.0 100.0 100.0 100.0 100.0 100.0 Number of respondents 710 54 87 80 69 Note: Source is MTM survey. Sample excludes respondents who died or were not traced in the final round. Table 3. Characteristics of Migrants and Non-Migrants at Baseline and Follow-up Men Women Stayer Moved Stayer Moved >1 Km >1 Km Characteristics at baseline: Age 20.3 20.5 16.6 17.0*** Relationship to head: Head 1.6 0.8 0.8 1.1 Child 76.5 58.7*** 78.9 66.5*** Grandchild 7.0 4.1 10.2 11.9 Niece or nephew 4.9 12.4*** 6.1 9.1* Sibling 4.9 11.6*** 1.9 1.7 Sibling-in-law 3.1 5.8* 1.9 4.6* Other 2.0 6.6*** 0.2 5.1*** Both parents alive (%) 70.8 62.8** 76.0 73.9 Household size 6.6 6.0** 6.3 6.3 Asset quartile of household: 19 Poorest quartile 25.9 20.2 26.2 23.4 25-50% 30.4 13.5*** 25.6 19.4* 50-75% 28.2 21.9* 26.7 24.6 *** Richest quartile 15.6 44.5 21.5 32.6*** Currently attending school (%) 43.4 33.1* 61.7 56.3 Main activity in last 12 months: Farming on household plots 38.8 29.6** 10.5 14.2 Non-farm income-generating work 24.3 28.9 4.4 4.0 ** Domestic chores 2.6 6.6 22.9 23.9 School 34.4 34.7 62.3 58.0 Savings for future (bank account, 11.9 9.9 3.6 6.8* savings group, cash, or assets) (%) Distance of village/community to Salima 17.4 13.8** 16.3 16.3 town (Kms) Characteristics at Follow-up: Currently married 25.1 28.9 25.1 55.7*** Relationship to head: Head/spouse of head 27.6 41.3*** 24.5 52.8*** Child 55.6 33.9*** 60.1 27.8*** Grandchild 6.5 0.0*** 7.7 1.1*** Niece or nephew 2.6 7.4*** 3.0 5.7* Sibling 4.9 8.3* 2.5 4.6* Sibling-in-law 0.8 4.1*** 1.7 3.4* Son/daughter-in-law 0.5 0.0* 0.0 4.0* Other 1.5 5.0* 0.5 0.6 * Both parents alive (%) 66.7 59.5 71.6 72.2 *** Household size 5.7 4.6 5.7 4.5*** Asset quartile of household: Poorest quartile 29.8 19.8** 24.3 17.6** 25-50% 27.7 18.2** 26.0 19.9* 50-75% 23.8 24.0 27.1 26.7 Richest quartile 18.7 38.0*** 22.7 35.8*** 20 Currently attending school (%) 16.7 13.7 34.7 19.8*** Main activity in last 12 months: Farming on household plots 36.4 31.6 19.7 21.5 Non-farm income-generating work 40.6 44.4 5.6 8.7* Domestic chores 3.4 8.6** 39.4 47.1** School 19.5 15.4 38.3 22.7*** Savings for future (bank account, 30.6 34.2 15.6 24.6*** savings group, cash, or assets) (%) Same household head as at baseline 62.2 20.7*** 67.1 13.1*** Number of respondents 387 121 363 176 Note: Source is MTM survey. Distance from baseline location is based on GPS data collected during the interview and not from self-reported distances. We test for the statistical difference between movers and non-movers, by gender. *** indicates significance at 1 percent; ** at 5 percent; and * at 10 percent. Table 4: Correlates of Migration after Baseline Moved >1 KM KMs moved Baseline characteristics (Probit, marginal (OLS) effects) Female 0.137*** 4.651 (0.045) (3.676) Age 0.009 0.844 (0.009) (0.760) Attending school -0.062** -1.321 (0.028) (1.548) Child of household head -0.103*** -4.092 (0.035) (2.703) Both parents alive 0.005 -0.235 (0.037) (3.000) Household size -0.014* -0.852* (0.007) (0.503) Distance of community/village 0.001 0.035 to Salima town (kms) (0.001) (0.076) Wealth index based on assets 0.091*** 9.413*** (0.017) (1.437) Number of observations 1,032 1,032 Note: *** indicates significance at 1 percent; ** at 5 percent; and * at 10 percent. Robust standard errors in parentheses. 21 22