Child Health and Migrant Parents in South-East Asia: Risk and Resilience among Primary School-Aged Children

Background to the CHAMPSEA Project

This special issue presents findings from a major research project investigating child health and migrant parents in South-East Asia (CHAMPSEA). Its aim is to contribute to the debate about a potential ‘crisis of care’ in the region as increasing numbers of parents migrate overseas for work, leaving their children behind (Parreñas, 2003). The project examines outcomes for two age groups of children in four study countries: Indonesia, the Philippines, Thailand and Vietnam. Here we focus on primary school-aged children of 9, 10 and 11 years both because this group tends to be neglected in the current literature on parental migration (but see Battistela and Conaco, 1998; ECMI/AOS-Manila et al., 2004) and because these pre-teens may be most at risk of any negative consequences of parental absence if they experience a care deficit. As the articles demonstrate, however, there is no simple distribution of risk and resilience between children living with both parents and the children of migrant parents. Rather there are important differences between countries and significant heterogeneity across multiple dimensions of health and well-being.

The international movement of labor from and within the global south is an increasing trend in many regions of the world, including South-East Asia. Motivations for migration may vary but, for parents of young children, working overseas is typically part of a household strategy aimed at securing an economic future, often by investing in their children’s education (Taylor, 1999). Both nations and households have become financially dependent on the remittances sent to family members who stay behind. While gains may be measured in terms of higher wages and increased household wealth, the effects of parental absence on children left behind are under-researched and therefore less clear. On the one hand, children may benefit if remittance monies are spent to improve their well-being and life chances. On the other hand, children may suffer if the absence of a parent results in a lack of social support and guidance, or causes psychological distress. The CHAMPSEA Project is designed to address this knowledge gap.

Three observations provide essential background to the themes discussed in the articles. First, it is important to note that international labor migration, while not a new phenomenon (DFID, 2007), is not a unitary one either. Rather, it represents a range of responses to changing economic, legal and social processes operating at a variety of scales from the globalization of capitalism through the policies of states to the intimacies of the family and the aspirations of the individual. In contemporary South-East Asia, the majority of international migrants take up 2 to 3-year employment contracts overseas that allow them to reside in destination countries only on a temporary basis, even though many migrants complete successive contracts that cover a much longer time period. Table 1 provides figures on the annual deployment of workers overseas for the four study countries, giving an indication of the size of the outward migration flows, although an unknown number of these migrants are parents who leave children behind. One of the distinguishing features of short-term contract migration in Asia and the Middle East is that the possibility of family reunification in the host country is restricted and there is rarely a pathway to citizenship or permanent resettlement for either the migrant or his/her family. When a parent migrates, therefore, other family members remain behind, creating a transnational household within which relationships must be maintained across distance. The growing academic literature on transnational families sometimes avoids describing children in transnational families as ‘left-behind,’ preferring to refer to them as children who ‘stay behind.’ The articles in this special issue use both terms. Although we resist the negative connotations often associated with describing children as ‘left-behind,’ we maintain that this descriptor does capture the lack of choice for parents on fixed-term contracts and the reduced agency of school-aged children themselves.

TABLE 1.

Migration Profiles for Study Countries

	(a)	(b)	(c)	(d)	(e)
Country	Total Population mid-2013 (millions)	Percent of population Ages<15	Infant mortality rate (per 1,000 live births)	GDP per capita US$# for 2012	Documented overseas workers
					Annual Deployment		% Female
Indonesia	248.5	29	32	$3,557	435,219	450,601	68	56.8
					(2000)	(2012)	(2000)	(2012)
Philippines	96.2	33	22	$2,587	933,588	1,470,826	741	551
					(2004)	(2010)	(2004)	(2010)
Thailand	66.2	19	11	$5,480	158,0002	143,795	15	15.7
					(2002)	(2010)	(2002)	(2010)
Vietnam	89.7	24	16	$1,596	30,000	85,546	15	approx. 30
					(2000)	(2010)	(2000)	(2010)

^#Notes: GDP per capita is gross domestic product divided by midyear population. Data are in current US dollars.

¹Percentage applies to land-based new hires only

²Number of documented Thai workers abroad

Sources: Columns (a)-(c): Population Reference Bureau (2013) “2013 World Population Data Sheet.” Available at http://prb.org/pdf13/2013-population-data-sheet_eng.pdf, accessed on 4 November 2013.

Column (d): The World Bank (2013) “GDP per capita (current US$)/Data/Table.” Available at http://data.worldbank.org/indicator/NY.GDP.PCAP.CD, accessed on 4 November 2013.

Column (e): Indonesia: 2000: “Ford, M. (2001) “Indonesian Women as Export Commodity: Notes from Tanjung Pinang,” Labour and Management in Development, 2(5):1-9. 2012: Ministry of Manpower and Transmigration of the Republic of Indonesia.“Statistics on Indonesian Workers by Destination Country and Gender.” Available at http://pusdatinaker.balitfo.depnakertrans.go.id/katalog/xdownload.php?f=372, accessed on 30 July 2013.

Philippines: For 2004 and 2010: Philippine Overseas Employment Administration. “Compendium of OFW Statistics.” Available at <http://www.poea.gov.ph/stats/statistics.html>

Thailand: 2002: Department of Employment, Ministry of Labour 2010: Huguet, J.W. and A. Chamratrithirong, (2011) Thailand Migration Report 2011. IOM Thailand: Thailand. Available at http://www.un.or.th/documents/tmr-2011.pdf, accessed on 30 July 2013.

Vietnam: 2000: http://www.eldis.org/vfile/upload/1/document/0903/Dhaka_CP_7.pdf 2010: Consular Department, Ministry of Foreign Affairs of Viet Nam (2012) “Review of Vietnamese Migration Abroad.” Available at http://eeas.europa.eu/delegations/vietnam/documents/eu_vietnam/vn_migration_abroad_en.pdf, accessed on 30 July 2013.

Secondly, the feminization of migration from countries such as the Philippines and Indonesia increased during the 1990s, with women comprising 74 percent and 68 percent of the annual deployment of documented overseas workers from the Philippines and Indonesia, respectively, by the early 2000s (see Table 1). This increase has engendered particular anxieties because the continuing dominance of traditional gender roles means that care for dependent children and elders usually has to be reconfigured when a mother migrates. Thus, the absence of migrant mothers is often assumed to present greater challenges to left-behind family members than the absence of migrant fathers. Gender ideologies are equally influential in promoting public anxieties about the effects of separation on children. In the Philippines, for example, it is the migration of mothers rather than fathers that has fuelled worries about left-behind children becoming spend-thrift, delinquent, addicted to drugs, and emotionally scarred (Asis, 2006; ECMI/AOS-Manila et al., 2004). Moreover, the challenges for those who stay behind are likely to be faced in circumstances of some uncertainty. The period of separation may be open-ended, stretching from a couple of years to a decade or more. Long absences, combined with a lack of clarity about when the migrant might (be able to) return, are likely to impact on family relationships. Children who remain in the country of origin may be especially affected by the ‘ambiguous loss’ (Suárez-Orozco et al., 2002) of a migrant parent.

Thirdly, research on transnationalism and intimate relations has tended to focus on adults as the key social agents, leaving children largely invisible (Orellana et al., 2001). Thus children’s voices “are usually not heard in the development of policies and programs that are intended to benefit them” (ECMI/AOS-Manila et al., 2004: 8). Further, research attention has only relatively recently turned from the benefits of remittances for transnational households towards the costs of what Ehrenreich and Hochschild (2003:3) called the migrant’s “commute.” As well as restoring the perspective and agency of children to studies of transnational migration, a key aspiration of the CHAMPSEA Project is to provide a systematic understanding of the impacts (both positive and negative) of living in a transnational household on children’s health and well-being. The articles in this volume use quantitative analytical methods to examine risks and resilience in relation a range of well-being measures for primary school-aged children (including nutritional status, school performance, general happiness, alcohol consumption and smoking behavior) in order to add to the evidence base for assessing the impacts of parental migration on children left behind.

Understanding Risk and Resilience

As Mazzucato and Schans (2011: 705) observed, studies on whether left-behind children benefit from parental migration depend “not only on the outcomes that are studied (economic vs. psychological outcomes) but also on the characteristics of the parent and child.” Variability in outcomes is also a symptom of the fact that there are few large-scale quantitative studies that systematically collect comparable data as most studies on the well-being of left-behind children in transnational family arrangements tend to be small-scale, qualitatively driven studies (Mazzucato and Schans, 2011: 705). Even when quantitative methodologies with large sample sizes are used, cross-country comparisons are often difficult to achieve without well-crafted research strategies that systematically incorporate comparability as a design feature. For example, two studies published in the same issue of Asian Population Studies on a similar theme – the impact of parental migration on left-behind children’s schooling – yield interesting but not directly comparable results. By integrating the Matlab Health and Socioeconomic Survey data with the Health and Demographic Surveillance System, Kuhn’s (2006) study of children’s pace of schooling in Matlab, Bangladesh, found that parental migration has significant and positive effects on the schooling of left-behind children, even after controlling for the long-term effects of socio-economic status. Focusing on the school enrollment of left-behind children in Kanchanaburi, Thailand, Jampaklay (2006), who used a 2002 and 2003 longitudinal study, found that the effects of migration on school enrollment of children depends not only on the absent status of their parents but also on the duration of their absence. While the study confirmed the positive effect of remittances on schooling, it also concluded that after controlling for remittances and other factors, parental migration has negative effects on the school enrollment of children.

Given the multidimensionality of children’s well-being (which may include psychological, social, educational and health outcomes) as well as the limited nature of available evidence and the lack of cross-national comparability, we argue that an approach that may be helpful in making sense of the patchwork diversity of studies is to revisit conceptions of ‘risk’ and ‘resilience’ not as independent factors but in tandem, as a productive pathway to understanding how children’s well-being is shaped in the context of parental migration.

Several studies have identified children who are said to be ‘at risk’ by expressing concern about child delinquency and risk-taking behavior as a result of parental migration. In their study of cigarette smoking among rural adolescents in South China, Gao et al. (2013) found that parental migration in general had unfavorable effects on self-efficacy, which was a strong influencing factor for smoking. More specifically, they concluded that while paternal migration was protective for adolescent smoking, maternal migration increased the risk. Turning to alcohol consumption and internal migration in Thailand, Jampaklay et al. (2012) found that children living in one-parent migrant households were most likely to be involved in alcohol drinking, while those in both-parent migrant households had the smallest chance compared to their peers. Liu (2012) reported that there was a rise in sexual abuse among left-behind children in China, where there are an estimated 58 million left-behind children. In accounting for why these children were at risk of sexual abuse, Liu (2012) explained that about 70 percent of the left-behind children were under the caregiving responsibilities of their grandparents, but that the grandparents were reported to be less watchful over the children and more reluctant to teach them sex education, leaving the children uninformed and hence vulnerable to abuse. Additionally, migrant parents who contacted their children tended to show greater concern over the children’s academic performance as opposed to their emotional health.

The limits to these studies indicate that a fuller explanation of children’s risk-taking behavior will need to extend beyond an analysis of the social and demographic correlates of the children (as well as those of the parents, surrogate caregivers, household, etc.) prone to risk behaviors (or lifestyle risks) to understanding children’s risk perception. As perceptions of risk where the benefits are seen to be more valuable than the risk may translate into risk-taking behavior, it follows that “differences in perception of risks could also explain differences in individual risk behaviour, so that risks that are taken are either perceived as low risks or as associated with very beneficial outcomes” (Hayenhjelm, 2006: 190). Hayenhjelm (2006) usefully suggested that in situations of extreme vulnerability or poverty, four other interconnected explanatory factors may influence risk perception and risk taking: the lack of reasonable options; poor outset conditions such that refraining from taking any action becomes another form of risk taking; hope for change that will better present circumstances; and liability to dis-information, especially in situations where there is no access to significant information, or no means to verify available information. Such an approach recognizes that risk factors are not just individual characteristics but also ecological or “environmental hazards that increase children’s vulnerability to negative … outcomes” (Engle et al., 1996: 621).

Inasmuch as more work on the subjective and contextual contours of risk behaviors can usefully supplement the identification of social, economic, demographic and household characteristics associated with these behaviors, a more concerted effort to recognize and understand the nature of resilience in children living in transnational households – as a positive complement to risk – can open up new directions to analyzing children’s well-being that go beyond documenting a “pathology of disadvantage” (Engle et al., 1996: 621). Researchers working on resilience are interested in “why some children manage to come through situations of multiple risks apparently unscathed,” or why certain individuals possess the “predisposition to resist the potential negative consequences of … risk” (Engle et al., 1996: 622; see also Walsh, 1996). They point to three interrelated domains that can enhance protective conditions and potentially strengthen resilience: the individual, the family and the larger social context.

At the individual level, Asis (2006) highlighted left-behind children’s capacity for agency and independence, reporting that some children choose to hide their problems from migrant parents in order not to burden them and instead find their own ways of coping, such as praying to God, minding their younger siblings, and contributing to “home work” and “care work” as positive strategies. At the family level, Walsh (1996) introduced the concept of “family resilience” as a means of understanding family functioning in relation to psychosocial challenges, constraints and resources. At its core, the term signifies “a narrative coherence that assists members in making meaning of their crisis experience and builds collaboration, competence, and confidence in surmounting family challenges” (Walsh, 1996:261). In this vein, Bryant (2007:10) suggested that the social costs of parental migration in the context of South-East Asia can be mitigated by the involvement of the extended family in caregiving responsibilities. Support and assistance provided by extended family members can help fill the care deficit in the absence of migrant parents who would in turn be assured that their children will be cared for while they are away (see also Battistella and Conaco, 1998; Asis, 2006). Jordan and Graham’s (2012) work pointed out that children seem to be more sensitive to proximate influences, such as their caregiver’s mental health status, than to parental absence, underscoring the significant role that family members acting as surrogate carers play in fostering children’s well-being. Finally, as in the case of research on risk, scholars working with notions of resilience also point to the importance of locating protective factors in the larger social and environmental context. This suggests that building resilience is also associated with the presence of supportive community organizations for the children of transnational households. One of the lessons from the CHAMPSEA Project is that different actors, including government agencies, civil society, businesses and local communities, need to work together in creating effective legal and institutional frameworks and sustainable mechanisms to support migrant families and their children (see Lam et al., this volume).

CHAMPSEA Methodology

The CHAMPSEA Project is a cross-sectional baseline study employing a mixed methods approach to investigate the health and well-being of two age groups of children: (1) pre-school children aged 3, 4 and 5 years; and (2) primary/elementary school children aged 9, 10 and 11 years. Data were collected on children living in both transnational (migrant) and non-migrant households, with the latter used as a control or comparator group in the quantitative analyses. This special issue focuses on the second age group who were themselves interviewed as part of the CHAMPSEA Project. Their responses form the basis of the analyses presented in several of the articles in this volume.

The Sample

In 2008, survey data were collected for around 1,000 index children and their households in each of the four study countries – Indonesia, the Philippines, Thailand, and Vietnam. Follow-up in-depth interviews with a small sub-sample of carers and school-aged children were conducted in 2009. Available national databases on overseas migration for these countries do not include information on migrant parents and thus there was no sampling frame from which to draw a nationally representative sample. Additionally, we noted that high out-migration communities tended to be geographically clustered. We therefore adopted a three-stage flexible quota sampling strategy adapted from “sentinel site surveillance” methods, as used in public health studies (Byass et al., 2002), and based on the method advocated by Wilson et al. (2006) for use in developing countries. Crucial to the strategy was the specification of detailed protocols such that any future replication should produce a sample equivalent to the CHAMPSEA sample in all main characteristics. The first stage used in-country experts to identify two provinces within each study country with international out-migration rates higher than the national average. East Java and West Java in Indonesia, Laguna and Bulacan in the Philippines, Lampang and Udon Thani in Thailand, and Thai Binh and Hai Duong in Vietnam were selected as meeting the criterion. This was followed by the identification of high out-migration communities outside the major metropolitan areas within each province, using local knowledge. Communities were then selected to ensure diversity (i.e., both long-established and more recent out-migration communities; or, in the case of the Philippines where international migration was established in the chosen provinces, both rural and more urbanized communities). The final stage involved the recruitment of qualifying households within selected communities (villages in Indonesia, the Philippines (i.e., barangays) and Thailand, and communes in Vietnam). The result is a sample for each country that is both replicable and appropriate to the objectives of the CHAMPSEA Project.

Recruitment to the CHAMPSEA survey involved community-based screening to find and select eligible households and index children. Qualifying households had to contain at least one child in a qualifying age group (3, 4, and 5 years, or 9, 10, and 11 years) and to be either (i) a transnational household in which one or both parents had been working overseas for at least six months prior to interview, or (ii) a non-migrant household in which both parents had been usually resident at the same address as the index child for at least six months prior to interview. Separated or divorced couples, single parent households and households where one or both parents were internal migrants were excluded from the survey in order to target resources and reduce analytical complexity. Quotas specified minimum and maximum numbers of households within groups defined by child age and gender, and by migration status (transnational and non-migrant households). Only one index child was selected in each qualifying household. Where two children in the same household qualified for the study, the selection was made according to which group was most under-quota or, where groups were equally under-quota, by random selection. Although the samples are not nationally representative, they are of sufficient size to conduct comparative analyses. Around half of the 1,000 households interviewed in each country contained index children aged 9, 10 or 11 and Table 2 provides a breakdown of the sample for this age group.

TABLE 2.

Composition of the CHAMPSEA Sample: Number of Index Children Aged 9, 10 and 11 by Country and Household Type

		Transnational household
Study Country	Non-migrant household	Father migrant	Mother migrant	Both parents migrant	Total
Indonesia	248	81	150	34	513
Philippines	245	172	66	17	500
Thailand	256	249	2	14	521
Vietnam	230	87	170	24	511
Total	979	589	386	89	2,045

The Survey

Survey interviews were conducted in local languages with one or more members of the index child’s household, using three separate survey schedules:

(1) The Household Questionnaire was administered to a responsible adult in the household who identified themselves as able to answer questions on family background and history, and the management and allocation of financial resources within the household. It included a household roster which collected summary information on both resident and non-resident members of the household, as well as regular day visitors.

(2) The Carer Questionnaire was administered to the person who identified themselves as the primary caregiver for the index child. In some cases, the same individual answered both the household and carer surveys; in other cases, it was two different individuals. Although the focus of the questions in this questionnaire was on the index child, data were also collected on those providing care for the child.

(3) The Older Child Questionnaire was administered to index children aged 9, 10 and 11 years, who were the older age group of children in the CHAMPSEA survey. Children were asked questions about their family and their own circumstances, including a series of questions on health-related behavior. Children in transnational households were also asked about their migrant parent/s.

In addition to the three survey instruments, anthropometric measures of height and weight were also taken for all index children.

All survey instruments were compiled in English and then subjected to a rigorous process of translation and back-translation to ensure accuracy. Pilot studies in all four countries were used to test local understandings and, where necessary, adjustments were made. Standard training was given to interviewers, including training on World Health Organization (WHO) standards in the use of specialist equipment for taking measurements of height and weight. All interviewers were fluent in local languages. The CHAMPSEA research team was sensitive to ethical issues that might arise during the survey, especially in the interviews conducted with the index children. The purpose of the survey was explained to potential participants and their anonymity assured. It was made clear that participation was voluntary, there would be no repercussions from non-participation, and participants could refuse to answer particular questions or withdraw from the interview at any time without giving a reason. Verbal informed consent was obtained from all adults prior to interview, and verbal assent from index children. Every effort was made to put children at ease and to explain that this was not a test and that there were no right or wrong answers. Children were interviewed within sight of an adult in their household but, where possible, out of earshot. Interviews were conducted by two interviewers in the participants’ own homes. Ethics approval for the CHAMPSEA Project was obtained from institutions in all participating countries.¹

Data and Measures

The data collected are wide-ranging and cover migration histories as well as many aspects of child health and well-being, including measures that are well-established and validated in the literature. The articles in this special issue examine different health and well-being outcomes but also use some of the same variables in their analyses. To avoid repetition, we therefore outline the most important of these here.

1. Migrant-carer status² is a derived variable that divides the sample into four categories according to both the migration status of the household and, for transnational households, who is the migrant and who is the principal caregiver for the index child. The four categories are: non-migrant; father-migrant/mother-carer; mother-migrant/father-carer; parent(s)-migrant/other-carer. This is our main migration variable and is usually preferred in CHAMPSEA analyses because it summarizes the situation of the child without introducing problems of collinearity.

2. Child psychological difficulties is measured using data from the Strengths and Difficulties Questionnaire (SDQ), developed by Robert Goodman (Goodman, 1997) and validated as a screening tool for child mental health problems in Asian settings. The 25 items of the SDQ were completed by the index child’s principal caregiver and provide scores for five subscales and a summary scale. The Total Difficulties Score sums four of the five subscales (emotional symptoms, conduct problems, hyperactivity/inattention, and peer relationship problems). Possible values range from 0 to 40, and a score of 17 or above is taken as identifying likely cases of children with psychological difficulties.

3. Family functioning is measured using the Family APGAR, a rapid screening tool widely employed in clinical practice and research with families. Children aged 9, 10 and 11 completed the questionnaire covering five parameters of family functioning: Adaptability, Partnership, Growth, Affection and Resolve (APGAR). All items were measured on a 5-point scale (0 never, 1 hardly ever, 2 some of the time, 3 almost always, 4 always) and scored by summing the values for the five items. The composite score ranges from 0 to 20 and can be used as a continuous variable, with higher scores signifying greater satisfaction with family functioning, or as a binary variable indicating either good or poor family functioning, depending on the cut-off used. Some of the papers in this volume identify good family functioning as a score of 13 and above.

4. Carer mental health is assessed using the WHO recommended Self-Reporting Questionnaire (SRQ), a screening tool for psychiatric disturbance recommended for use in low income countries and now validated in many cultural contexts. All principal carers of the index children were asked 20 yes/no questions relating to symptoms of depression and anxiety in the previous 30 days. ‘Yes’ answers were summed to provide a score, and scores of 8 or more are regarded as indicating probable cases of poor mental health.

5. Anthropometric measurements (height and weight) were taken for all index children, together with exact age. These data allow the calculation of three measures of child undernutrition (wasting, stunting and underweight). The measure used in several articles in this volume is Child stunting, or low height-for-age. Height-for-age z-scores (HAZ) were calculated using the WHO Anthro software, and stunting is defined by a HAZ score more than two standard deviations below the median of the 2006 WHO Child Growth Standards.

6. Household wealth is measured using a wealth index, based on the methodology developed by the Young Lives study for use in low and medium income countries (see www.younglives.org.uk). The index combines data on housing quality (e.g., wall and roof material), consumer durables (e.g., television; working motor bike/scooter) and services (e.g., electricity; piped drinking water) to derive an overall score. Quintiles are used to classify households into low (first and second quintile), medium (third and fourth quintile) and high (fifth quintile) wealth categories. The classification can be derived for a single country, showing relative wealth within that setting, or for several countries together, showing relative wealth within the combined group. Care must be taken in interpreting the single-country measure when comparing across countries, as the measure of ‘high wealth’ in Vietnam is not directly comparable to ‘high wealth’ in the Philippines, for example.

Analytical Methods

The analyses in this special issue use quantitative analytical methods to compare outcomes for children living in different types of transnational household with outcomes for a ‘control group,’ namely children living with both parents in non-migrant households in the same communities. Methods of multivariate regression analysis are used, as appropriate to the cross-sectional nature of the CHAMPSEA data. These and other statistical tests are described in the individual articles.

Introducing the Articles

The special issue presents five articles and a research note, each investigating a different aspect of parental migration and its impact on children left behind. Together, they provide a glimpse of the richness of the CHAMPSEA dataset. The analyses exploit only part of that dataset, however, as they utilize the survey data for primary school-aged children rather than all index children, and concentrate on measurable health and well-being outcomes. The overarching question is: What difference does parental migration make?

In the first article, “Does Having a Migrant Parent Reduce the Risk of Undernutrition for Children Who Stay Behind in South-East Asia?,” Elspeth Graham and Lucy P. Jordan investigate the impact of parental migration on child nutrition. They use a standard measure of nutritional status that identifies low height-for-age (i.e., stunting) as their outcome of interest. Despite the extensive literature on child stunting and attempts to reduce its high prevalence in Africa and Asia to meet the Millennium Development Goals, there have been no previous studies of the potential impact of parental migration on the likelihood of a child being, or remaining, stunted. The article examines child stunting in the Philippines and Vietnam, comparing children in different types of transnational household with children in the same communities living with both parents. The findings from a series of multivariate models are interesting because they challenge any general assumption that left-behind children have better nutrition than their peers due to the higher earnings of their parents working abroad. A more nuanced picture emerges in which a reduced risk of stunting is apparent only for children of migrant fathers left in the care of their mothers in the Philippines. Having a migrant father seems to have no effect in Vietnam. In neither country is having a migrant mother associated with a lower risk of stunting compared with children in non-migrant households, but nor is the risk higher. Thus, there is no evidence in this analysis to support popular notions of a ‘crisis of care’ when mothers seek work in another country.

The great majority of migrant parents send money home to support family members who stay behind. Some transnational households in the CHAMPSEA Project used these remittances to buy everyday necessities, including food. Many parents, however, directed these overseas earnings towards improving the longer-term life chances for their children by investing in their schooling. Indeed the follow-up qualitative interviews with the children revealed that education was one of the main topics of conversation in communications with their migrant parents. In transnational households, therefore, particular emphasis may be placed on doing well at school, but how much difference does it make? The second article, “Leaving a Legacy: Parental Migration and School Outcomes among Young Children in the Philippines” by Maruja M.B. Asis and Cecilia Ruiz-Marave, examines the evidence for the Philippines. The article adds to the limited literature on schooling and left-behind children by investigating school pacing, or progression, and school achievement. The findings suggest that having a migrant parent is an advantage for some children when it comes to both school pacing and school achievement. In particular, children of migrant fathers looked after by their mothers are more likely to be at pace or advanced and to have higher school achievement compared to children in non-migrant households. There is no evidence of any disadvantage for the children of migrant mothers looked after by their fathers (or other carers) when compared to their peers living with both parents. Thus, although only some children appear to benefit from having a migrant parent, the concern that parental absence due to migration can negatively affect the school performance of children is not supported by this study.

The theme of resilience, as opposed to risk, provides a focus for the third article on Thailand, “The Subjective Well-Being of Children in Transnational and Non-migrant Households: Evidence from Thailand” by Aree Jampaklay and Patama Vapattanawong. Children’s subjective well-being is investigated to discover which groups of children regard themselves as ‘doing well,’ defined in this article with respect to how happy children say they are and how much they enjoy school. Contrary to the expectations of some, the findings show that children living in transnational households are more likely to be resilient on this measure than children living in non-migrant households. In the Thai case, however, the comparison is between left-behind children of migrant fathers and children in non-migrant households because so few mothers migrate leaving their children behind. Nevertheless, it is children living with both parents who are found to be relatively disadvantaged and those living in transnational households who appear to be more resilient.

While having a migrant parent may be an advantage in some respects, an important worry associated with the ‘crisis of care’ in the popular imagination is that children living in transnational households may lack guidance and may thus be more likely to be involved in risky health-related behaviors such as drinking and smoking. The fourth article, “Alcohol Use among Very Early Adolescents in Vietnam: What Difference Does Parental Migration Make?” by Lucy P. Jordan, Elspeth Graham and Duc Vinh Nguyen and the research note at the end of the collection, “Tobacco Use and Exposure among Children in Migrant and Non-migrant Households in Java, Indonesia,” by Sukamdi and Anna Marie Wattie, investigate the associations between parental migration and, respectively, alcohol use among Vietnamese children and tobacco use among Indonesian children. Even at 9, 10 and 11 years old, there were a number of children in the CHAMPSEA samples who admitted that they had drunk alcohol or smoked tobacco, with prevalence rates of 16.2 percent for alcohol use in Vietnam and 8.6 percent for tobacco use in Indonesia. In both studies, boys were found to be significantly more likely to participate in these behaviors than girls, and the factor most strongly associated with the behaviors was whether friends used alcohol (Vietnam) or tobacco (Indonesia). The health-related behaviors of other family members had little or no impact, and no evidence was found to support the idea that children of migrant parents are more at risk than children in non-migrant households. Indeed, the study in Vietnam reported significantly lower odds of alcohol use among children in transnational households left in the care of their father or someone other than a parent, which in most cases would be a grandparent.

The special issue is rounded off with an article, “Securing a Better Living Environment for Left-Behind Children: Implications and Challenges for Policies” by Theodora Lam, Miriam Ee, Hoang Lan Anh and Brenda S.A. Yeoh, that considers the challenges and policy implications of the CHAMPSEA study. While left-behind children show considerable resilience, they and their families are also facing a number of risks. Areas of concern for migration and development policy include the provision of better support for left-behind caregivers themselves as they substitute for absent migrants in providing care work, improvements to communication infrastructure to help migrants and their families maintain their relationships across transnational spaces, and assistance to migrant families under considerable stress stemming from the cycle of debts resulting from debt-financed migration. In view of these concerns, the article urges governments to collaborate with civil society and businesses to create effective legal and institutional frameworks as well as suitable supporting mechanisms for the growing population of left-behind families and their children. These measures are all the more urgent given the current thinking that migration can be part of a sustainable development strategy to ensure upward socio-economic mobility for Southeast Asian families.