1、<p>　　字?jǐn)?shù)：英文3980單詞，21223字符；中文7803漢字</p><p>　　出處：Nguyen C V. Does parental migration really benefit left-behind children? Comparative evidence from Ethiopia, India, Peru and Vietnam[J]. Social Science &am

2、p; Medicine, 2016, 153: 230-239.</p><p><b>　　外文文獻(xiàn)　</b></p><p>　　Does parental migration really benefit left-behind children? Comparative evidence from Ethiopia, India, Peru and Vietn

3、am</p><p>　　Abstract Millions of children are left behind when their parents migrate from home to another place. This study examines whether parental migration can affect health and cognitive ability of lef

4、t-behind children aged at 5e8 years old in Ethiopia, India, Peru, and Vietnam. It uses data on 7725 children in the four countries collected from Young Lives surveys in 2007 and 2009. It finds that although parental migr

5、ation helps families increase per capita consumption, it does not improve health and </p><p>　　Keywords: Ethiopia; India; Peru; Vietnam; Child welfare; Child nutrition; Parental migration; Parental absence&l

6、t;/p><p>　　1.Introduction</p><p>　　An important trend associated with economic and medical development is improvement in children's nutrition and health. Yet, it is estimated that 23.8 percent

7、and 7.5 percent of under-five children were still stunted and wasted in 2014, equivalent to around 159 and 50 million children, respectively (UNICEF, WHO, and World Bank, 2015). Most the malnourished children are living

8、in developing countries in Asia and Africa. At the same time, recent concern about health of children is a rising rate of </p><p>　　Health improvement for children has received a great amount of attention fr

9、om researchers and policy makers. Social and medicine science has long emphasized the important role of early childhood investments in improving health, education and return to education (e.g., Victora et al., 2008; Camp

10、bell et al., 2014; Heckman, 2012). Regarding nutrition for children, there is an influential argument that nutrition during the 1000 days consisting of the maternal period and 24 months after birth plays t</p><

11、;p>　　In this study, we will examine whether health and cognitive skills of children after the 1000 day window can be still affected by an important factor: parental migration. More specifically, we will estimate the i

12、mpact of parental migration, both short-term and long-term, on health and cognitive skills of children aged 5e8 in four countries including Ethiopia, India, Peru and Vietnam. Migration is an important livelihood strategy

13、 of households. There has been an increasing trend of parental migra</p><p>　　The main problem of parental migration is that its total effect on left-behind children is a priori unknown. On the one hand, par

14、ental migration can increase household income through remittances. Increased income can lead to in an increase in spending on food, health and education for children (e.g., Adams and Page, 2005; Basu et al., 2010). Howev

15、er, on the other hand, parental migration means parental absence at home, and as a result less attention and care for children, and a negative effect o</p><p>　　Empirical evidence on the impact of migration

16、on children is also mixed. Several studies find that migration and remittances help children increase education and reduce child labor (e.g., Yang, 2008; Antman, 2010; Alcaraz et al., 2012; Binci and Giannelli, 2012). Mi

17、gration is also found to improve health and nutrition of children (Hildebrandt and McKenzie, 2005; Macours and Vakis, 2010; Anton, 2010). Nevertheless, there are a number of studies which find negative effects of parenta

18、l migration on</p><p>　　In this study, we examine the effect of parental migration on health (measured by anthropometric indicators) and cognitive skills of children in four developing countries in three con

19、tinents including Ethiopia, India, Peru and Vietnam using panel data from the Young Lives surveys in 2007 and 2009. The availability of panel data allows us to apply a child fixed-effects estimator to estimate the impact

20、 of parental migration. The main advantage of the fixed-effects estimator is that it can control </p><p>　　The study is expected to have two contributions to the literature of child nutrition and migration.

21、Firstly, this study shows that children's health after five years old can be still affected by parental migration even after the health status at infancy is controlled for. Secondly, this study provides findings on t

22、he effect of migration on not only health but also cognitive skills of children, and comparable evidence on different countries and different types of parental migration including shor</p><p>　　2.Economic th

23、eories and literature review</p><p>　　Nutrition is essential for growth and development of human capital. Nutrition interventions have been implemented with focus on the duration of women's pregnancy and

24、 24 months after birth (UNICEF, 2013). The 1000 day window is considered as the critical period for accumulation of health and nutrition during this period can affect health and productivity of adults strongly (e.g., Mar

25、torell et al., 1994; Victora et al., 2008, 2010; Martorell et al., 2010; Hoddinott et al., 2008; Maluccio et al., 2</p><p>　　People can migrate for different incentives, but mainly for economic ones (Stark a

26、nd Bloom, 1985; and Stark, 1991). Households will migrate or send migrants if the migration benefit is larger than the migration cost. A common trend in migration is that either one of parents or both parents migrate for

27、 work and leave their children behind. Migration of parents can affect the large number of children's outcomes including consumption, education, and health through several channels. The most direct </p><p&

28、gt;　　Empirically, there are a number of studies finding that migration and remittances help children increase education enrollment or attainment. Hanson (2002), Antman (2010) and Alcaraz et al. (2012) find that internati

29、onal migration and remittances help children in Mexico increase education and reduce child labor. Yang (2008) finds a similar effect of international remittances on child education and labor in Philippines. Binci and Gia

30、nnelli (2012) focus on the impact of migration on the well-being </p><p>　　Regarding health outcomes, Hummer (2002) finds a lower proportion of Mexican children with low birth weight in migrating households.

31、 Hildebrandt and McKenzie (2005) examine the impact of international migration on child health outcomes in Mexico and find that children in migrating households have lower rates of infant mortality and higher birth weigh

32、ts. In Carletto et al. (2011), nutrition of children in migrant-sending households in Guatemala is more improved than children in households without</p><p>　　However, parental migration does not always benef

33、it leftbehind children for at least two reasons. Firstly, migration does not necessarily lead to higher income of migrant-sending households. Migration can also prevent these households from highreturn but labor intensiv

34、e activities (Taylor and Lopez-Feldman, 2010). Moreover, remittances can reduce work incentives for recipients (Farrington and Slater, 2006; Sahn and Alderman, 1996). Secondly, with less care from parents due to absence,

35、 children m</p><p>　　A large number of studies conclude negative effects of parental migration on children's education (e.g., McKenzie and Rapoport, 2006; Lahaie et al., 2009; Giannelli and Mangiavacchi,

36、 2010; Robles et al., 2011; Wang, 2011; Coffey, 2013; Lu, 2014; Zhang et al., 2014). Adverse impacts of parental migration on children's health are also found in several empirical studies. For example, Kiros and Whit

37、e (2004) find that children in Ethiopia with migrating mothers tend to have less immunisation coverage</p><p>　　3.Data set</p><p>　　In this study, we use data from the Young Lives project, which

38、 is an international study of childhood poverty collecting data on 12,000 children and their families in Ethiopia, India, Peru and Vietnam over 15 years. The project is led by a team in the Department of International De

39、velopment at the University of Oxford, the United Kingdom with funding from the Department for International Development (DFID), the Netherlands Ministry of Foreign Affairs, and Irish Aid.</p><p>　　There are

40、 two groups of children surveyed in the Young Lives project: (i) the Younger Cohort, consisting of 2000 children who were born in 2001 and 2002, and (ii) the Older Cohort, consisting of 1000 children who were born in 199

41、4 and 1995. The sampled children are representative of the poor areas instead of the national level. In each of the four countries, 20 sites (the sites are equivalent to districts) are selected from poor areas. In each s

42、ite, enumeration areas such as communes or villages</p><p>　　Up to now, three rounds of the surveys have been completed: in 2002, 2007 and 2009. In this study, we will use data from the younger cohort in the

43、 second round (when they were 5) and the third round (when they were 8) of the Young Lives surveys. We do not use the data in the first round, since there is no information on children's education in this round. Data

44、 from the older cohort are not used either, since there are no data on parental migration.</p><p>　　Initially, in the first round the number of children in the sample of each country is 2000. However, becaus

45、e of attrition, the number of children in the younger cohort is 1914, 1950, 1963, and 1970 in the samples of Ethiopia, India, Peru and Vietnam in the second round survey, respectively. In the third round, the respective

46、number of children is 1885, 1930, 1946 and 1964.</p><p>　　One of important issues in migration studies is how to define migration. According to Lipton (1980), migration is referred as “the departure of indiv

47、iduals or households, for more than a week or so” from their home. In some studies, a household member is defined as a migrant if she/he is absent from home more than 6 months (e.g., De Brauw and Mu, 2011; and Nguyen et

48、al., 2011). A member is defined as a short-term migrant if she/he departs from home several durations during the years (De Brauw an</p><p>　　There are no sections on migration in Young Lives surveys. However

49、, we can use information from a parental background section to define parental migration. More specifically, there is a question on ‘How often do parents see the child’. If parents stay home with children, the answer mus

50、t be ‘daily’. Parents who work far from home would ‘see their children’ less often, that is weekly, monthly, annually and less. To define migration, we use this information and also information on working and divor</p

51、><p>　　It should be noted that in this study the treatment is defined as parents who works far from home and do not see their children on the daily basis. Children who have divorced or death parents are not con

52、sidered. The control group includes children who have both mothers and fathers working around home. We do not consider the effect of migration of the whole family.</p><p>　　4.Parental migration and children&

53、#39;s outcomes</p><p>　　4.1.Parental migration</p><p>　　This section presents the descriptive analysis of parental migration and children's outcome in Ethiopia, India, Peru and Vietnam. Amon

54、g these countries, Peru is the richest country, while Ethiopia is the poorest. India and Vietnam have a similar economic level. In term of area and population, India is the largest country with the highest population (Ta

55、ble 1).</p><p>　　Table 1 Country background in 2012.</p><p>　　Table 2 examines the parental migration rate by estimating the percentage of children seeing their mother and father at particular i

56、ntervals. As mentioned, we define a father or a mother as a migrant if he or she does not see their children daily. Table 2 shows several noting points. Firstly, in the four countries, children were more likely to see mo

57、thers than fathers. Children also had a higher rate of divorced or death fathers than mothers. In developing countries, women will find it more diff</p><p>　　Table 2 Parental migration in 2007 and 2009.</

58、p><p>　　Secondly, the rate of short-term migration which is defined by ‘weekly seeing’ and ‘monthly seeing’ tended to be higher than the rate of long-term migration which is defined by ‘a(chǎn)nnually seeing’ and ‘ir

59、regularly seeing’. The long-term migration rate was higher in Ethiopia and Peru than in Vietnam and India.</p><p>　　Thirdly, the migration rate of parents varied across the four countries, and there was a la

60、rge gap in parental migration between these countries (Fig. 1). For example, in 2009, the rate of mothers' migration in India was nearly double the rate in Ethiopia. Peru was the country having the highest rate of fa

61、thers' migration, followed by India, Ethiopia and Vietnam. In 2009, the rate of fathers' migration in Peru was twice as much as the rate in other three countries.</p><p>　　Fig. 1. The percentage of c

62、hildren having parental migration. </p><p>　　Over the period 2007-2009, more children had mothers' migration. Regarding fathers' migration, this rate increased in India and Peru, but slightly decreas

63、ed in Ethiopia and Vietnam.</p><p>　　Table 3 presents the migration rate in urban and rural areas in both years. In Ethiopia and Peru, urban parents are more likely to migrate than rural parents. However, in

64、 India and Vietnam, the trend is opposite: rural people were more likely to migrate than urban people. The association between education and migration differs for different countries. In India and Peru, parents with prim

65、ary educational degree or less were more likely to work far from home than those with higher educational degre</p><p>　　Table 3 Parental migration by urban status and education levels.</p><p>　　

66、4.2.Children outcomes</p><p>　　In this study, we measure the health status of children by weight-for-age z-score and height-for-age z-score. These variables are also widely used in studies on the impact of m

67、igration on children's health (e.g., Macours and Vakis, 2010; Anton, 2010; Langworthy, 2011; Gibson et al., 2011). In the Young Lives data, zscore is computed as the ratio of the difference between child's measur

68、ement and the median to the standard deviation of the children in an international reference (standard) group de</p><p>　　Table 4 Nutrition of children in 2007 and 2009.</p><p>　　Children with Z

69、-score below —2 can be considered as malnourished (WHO, 2007). Peru has the lowest proportion of children with weight-for-age z-score below —2, at 3.5% in 2007 and 5.3% in 2009. This rate is very high in India, nearly 40

70、% in both 2007 and 2009. In terms of height, the proportion of children with height-for-age z-score below —2 is quite similar in the four countries.</p><p>　　Recently, attention has been paid to not only mal

71、nutrition but also obesity of children. Body Mass Index (BMI) for age can be used to examine overweight and obesity of children. According to WHO (2007), a child can be considered as obese and overweight if her/his BMI-f

72、or-age Z score is above 1 and 2, respectively. Table 4 shows that the proportion of overweight and obese children was highest in Peru, followed by Vietnam, then Ethiopia and India.</p><p>　　In Ethiopia, Indi

73、a and Peru, children with parental migration tended to have higher weight and height than children without parental migration. The proportion of malnourished children (i.e., children with Z-score below —2) was also lower

74、 in families with parental migration than families without parental migration. However, in Peru, children with parental migration had a higher rate of overweight and obesity than those without parental migration. In Viet

75、nam, children with and children without par</p><p>　　To measure cognitive ability of children, we use two test scores. The first is Peabody Picture Vocabulary Test (PPVT), which is a widely-used test of rece

76、ptive vocabulary. Interviewers present a series of pictures to each child. For each item, there are four pictures in a page, and each is numbered. Interviewers describe one of the pictures verbally and ask the child to p

77、ick up a picture that is described. Several studies find that the PPVT has a positive strong correlation with some commonly us</p><p>　　The second is a math test, which is designed based on the publicly rele

78、ased items of the Trends in International Mathematics and Science Study developed by the International Evaluation Association in 2003. The test focuses on addition, subtraction, multiplication, division. In the 2009 surv

79、ey, the math test includes 20 items dealing with addition, subtraction, multiplication, division and square roots, using both whole numbers and fractions, and 10 items on mathematics problem solving. The range</p>

80、<p>　　In the 2007 surveys, children were not attending primary school. As a result, a so-called a Cognitive Development Assessment (CDA) test instead of a math test was applied to the children. This test also measu

81、res the cognitive ability of children in a quantitative way. The test includes simple questions about spatial relations, quantity, and time. There are 15 questions and the test score ranges from 0 to 15. Table 5 presents

82、 the summary statistics of the PPVT, math and CDA test scores in the 200</p><p>　　Table 5 Cognitive ability tests and per capita expenditure in 2007 and 2009.</p><p>　　Finally, we look at per ca

83、pita expenditure of children's families. It shows that the real per capita expenditure increased overtime, and families with parental migration tended to have higher per capita expenditure than families without paren

84、tal migration.</p><p>　　7.Conclusions</p><p>　　This study examines the effect of parental migration on health and cognitive skills of children aged at 8 years old in Ethiopia, India, Peru, and V

85、ietnam. We find that parental migration, especially long-term migration, helps families increase per capita expenditure in India and Vietnam. However, parental migration is not beneficial for children's health and ed

86、ucation. Although, we do not find a significant effect of parental migration on children in Ethiopia, we a negative effect of parental mi</p><p>　　Our findings suggest several policy implications. Firstly, h

87、ealth and nutrition of children can be still affected after five years old. Attention from parents is an important factor to improve health of children. Secondly, although parental migration can be an important way to in

88、crease household income, it can have harmful effects on children left behind. Nutrition support programs should be provided for not only poor children but also those with parental absence. Thirdly, frequent contacts betw

89、</p><p><b>　　中文譯文</b></p><p>　　父母外出務(wù)工真的有利于留守兒童嗎?</p><p>　　來自埃塞俄比亞、印度、秘魯和越南的對比證據(jù)</p><p>　　摘要 當(dāng)父母從家外出務(wù)工到另一個地方時，上百萬的孩子被留在家里。本研究考察了父母外出務(wù)工是否會影響埃塞俄比亞、印度、秘魯和越南的5歲

90、至8歲的留守兒童的健康和認(rèn)知能力。本文使用了2007年和2009年從Young Lives調(diào)查中收集的四個國家7725名兒童的數(shù)據(jù)。研究發(fā)現(xiàn)，盡管父母外出務(wù)工有助于家庭增加人均消費，但并未改善兒童的健康和認(rèn)知能力。父母外出務(wù)工的影響因不同國家和不同類型的外出務(wù)工而異。在埃塞俄比亞，父母外出務(wù)工對兒童沒有顯著影響。然而，父母外出務(wù)工會降低其他三個國家兒童的健康狀況，并降低印度和越南兒童的認(rèn)知能力測驗分?jǐn)?shù)。長期父母外出務(wù)工對兒童的負(fù)面影響往

91、往高于短期父母外出務(wù)工。</p><p>　　關(guān)鍵詞：埃塞俄比亞；印度；秘魯；越南；兒童福利；兒童營養(yǎng)；父母外出務(wù)工；父母缺席</p><p><b>　　1.簡介</b></p><p>　　與經(jīng)濟(jì)和醫(yī)學(xué)發(fā)展有關(guān)的一個重要趨勢是改善兒童的營養(yǎng)和健康。然而，據(jù)估計，2014年仍有23.8%和7.5%的五歲以下兒童發(fā)育遲緩和瘦骨嶙峋，分別相當(dāng)于約

92、1.59億和5000萬兒童（聯(lián)合國兒童基金會，世衛(wèi)組織和世界銀行，2015）。大多數(shù)營養(yǎng)不良的兒童生活在亞洲和非洲的發(fā)展中國家。與此同時，最近對兒童健康的關(guān)注是超重和肥胖率的上升。據(jù)估計，2014年全球有6.1%的兒童（約4100萬）超重（聯(lián)合國兒童基金會，世衛(wèi)組織和世界銀行，2015）。</p><p>　　改善兒童健康狀況受到了研究人員和決策者的極大關(guān)注。社會和醫(yī)學(xué)科學(xué)一直強(qiáng)調(diào)兒童早期投資在改善健康、教育和重

93、返教育方面的重要作用（例如，Victora等，2008；Campbell等，2014；Heckman，2012）。關(guān)于兒童的營養(yǎng)問題，有一種有影響力的觀點認(rèn)為，即孕產(chǎn)期和出生后24個月的1000天營養(yǎng)對兒童的健康乃至成人的生產(chǎn)力起著關(guān)鍵作用（例如，Martorell等，1994；Victora等，2008，2010；Martorell等，2010；Hoddinott等，2008；Maluccio等，2009）。在這個關(guān)鍵時期造成的營養(yǎng)缺

94、乏在大約2歲后很難逆轉(zhuǎn)。然而，其他研究發(fā)現(xiàn)，嬰兒時期過后，健康狀況可能會有相當(dāng)大的生長恢復(fù)（例如，Lundeen等，2014；Prentice等，2013；Mani，2012）。Crookston等人（2013）的研究表明，兩歲后健康狀況的改善可以幫助兒童提高數(shù)學(xué)和閱讀成績。家庭變量，如消費支出和父母教育，可以影響營養(yǎng)在嬰兒期和后期之間的流動（Schott等，2013）。</p><p>　　在這項研究中，我們將

95、檢查1000天后兒童的健康和認(rèn)知技能是否仍然受到一個重要因素的影響——父母的外出務(wù)工。更具體地說，我們將評估埃塞俄比亞、印度、秘魯和越南等四個國家的父母短期和長期外出務(wù)工對5歲至8歲兒童的健康和認(rèn)知技能的影響。外出務(wù)工是家庭的一項重要生計戰(zhàn)略。父母外出務(wù)工的趨勢日益明顯，特別是在發(fā)展中國家。然而，隨著父母的外出務(wù)工，數(shù)以百萬計的兒童與單親父母或祖父母呆在家里。人們越來越關(guān)注父母外出務(wù)工對留守兒童可能產(chǎn)生的不利影響。</p>

96、<p>　　父母外出務(wù)工的主要問題是，它對留守兒童的總體影響是未知的。一方面，父母外出務(wù)工可以通過匯款增加家庭收入。收入的增加可促進(jìn)兒童食品、保健和教育支出的增加（例如，Adams和Page，2005；Basu等，2010）。然而，另一方面，父母外出務(wù)工意味著父母不在家，因此對兒童的關(guān)注和照顧較少，對兒童的健康和教育產(chǎn)生不利影響。此外，兒童可能不得不工作或照顧其他受扶養(yǎng)的成員，花在教育和休息上的時間更少（McKenzie和R

97、apoport，2006；Cappelloni，2011；Antman，2012）。</p><p>　　關(guān)于外出務(wù)工對兒童影響的經(jīng)驗證據(jù)也參差不齊。一些研究發(fā)現(xiàn)，外出務(wù)工和匯款有助于兒童增加教育并減少童工（例如，Yang，2008；Antman，2010；Alcaraz等，2012；Binci和Giannelli，2012）。外出務(wù)工還有助于改善兒童的健康和營養(yǎng)（Hildebrandt和McKenzie，200