Poverty, Risk and Families' Responses: Evidence from Young Lives more

Policy Paper 4 Poverty, Risk and Families’ Responses: Evidence from Young Lives September 2011 Kirrily Pells Acknowledgements I am especially grateful to Laura Valadez and Daniel Vujcich for assistance with data analysis and to Paul Dornan for assistance with the development of the paper. Also to Inka Barnett and Jo Boyden for helpful comments. My thanks are due particularly to the data-gathering teams in Ethiopia, Andhra Pradesh (India), Peru and Vietnam who carried out the data collection, upon which the report is written. Readers are encouraged to quote or reproduce material from Young Lives papers in their own publications. In return, Young Lives requests due acknowledgement and a copy of the publication. Young Lives is a 15-year study of childhood poverty in Ethiopia, the state of Andhra Pradesh in India, Peru and Vietnam, following the lives of 3,000 children in each country. It is core-funded from 2001 to 2017 by UK aid from the Department for International Development (DFID) and co-funded by the Netherlands Ministry of Foreign Affairs from 2010 to 2014. The full text of all Young Lives publications and more information about our work is available on our website. http://www.younglives.org.uk © Young Lives, 2011 The Author Kirrily Pells is Policy Officer at Young Lives. She completed a PhD focusing on rights-based approaches with children and young people in post-conflict situations, with a case study on Rwanda. This was followed by a postdoctoral fellowship at the School of Advanced Study, University of London. She has been a consultant for Save the Children UK, CARE International and the UK Foreign and Commonwealth Office. She is currently focusing on gender, risk and resilience and linking research with policy and practice. Young Lives Policy Paper 4: September 2011 Page 1 Contents Summary and findings 1. 2. 3. 4. Introduction About Young Lives Risk, resilience and protection Disparities in life chances widen over the life course 4.1 4.2 Risks can be cumulative and compounded for poor children Chronic poverty and inequalities repeatedly disadvantage the same children 2 4 5 5 7 10 14 5. Poverty and inequalities are at the root of risks for children 5.1 5.2 5.3 5.4 Poor households are affected by multiple and recurrent experiences of illness Illness is a cause and consequence of poverty Children and households depend on multiple sources of informal and formal social support in responding to illness Access to quality healthcare can strengthen protection for children 18 18 20 22 25 6. Conclusions and policy discussion: from ‘beating the odds’ to ‘changing the odds’ 29 30 References www.younglives.org.uk/what-we-do/our-policy-work Page 2 Poverty, Risk and Families’ Responses Summary and findings This paper brings together existing Young Lives research and policy analysis, alongside new findings, to argue that poverty and inequalities are at the heart of childhood risk, shaping which children are at risk, access to sources of protection, and children’s life chances. Drawing on the rounds of survey and qualitative data collection conducted to date, it illustrates how risk is mediated through poverty and structural disadvantage, meaning that children from groups with low social status, from rural areas and the poorest households, have increased risk of having poorer outcomes in education, health and subjective well-being indicators. Policymakers concerned with reducing risk and improving protection should not focus on enabling individual children to ‘beat the odds’ but instead on ‘changing the odds’ (Seccombe 2002). This means targeting the root causes of children’s poor life chances, namely poverty and inequalities, rather than just the symptoms of risk. 1. Disparities in life chances widen over the life course. As the children become older the difference in school enrolment and cognitive achievement between non-poor children and poor children, between children living in urban and rural areas, and between groups with different social status because of ethnicity or caste, increase over children’s lives. Household wealth is a significant predictor of continued school enrolment and cognitive achievement. For example, in Vietnam the percentage point difference in school enrolment between poor children and the rest more than tripled between the ages of 12 and 15. 2. Risk factors can be cumulative and compounded but often in more nuanced ways than sometimes assumed in advocacy initiatives. Taking gender as an example, differences between boys and girls are often smaller than other disparities, yet gender differences appear to become more significant as children age. Higher maternal education has an independent effect additional to consumption or wealth, associated with higher child cognitive achievement across all four study countries. 3. Chronic poverty and inequalities repeatedly disadvantage the same children. Children from ethnic minorities/low-caste groups and from rural and the poorest households are over-represented among children who have an increased risk of doing less well across a series of indicators in education, health and subjective well-being. This suggests that risk is mediated through poverty and structural disadvantage. 4. Children and families living in chronic poverty carry a higher burden of risk, as they have fewer resources to cope with risks or adverse events such as illness and death, meaning that they are hit harder, which in turn can perpetuate the cycle of poverty and risk. The entire household, including children, are involved in managing ill-health and responding to death, both of which are considerable sources of anxiety for poor families and can have long-term consequences for children’s life chances. For example, in Andhra Pradesh children who have lost a family member (other than a caregiver) are 70 per cent less likely to be enrolled in school. In both Ethiopia and Vietnam the death of a father is associated with lower cognitive achievement, perhaps suggesting periods of absence or drop-out from school. www.younglives.org.uk/what-we-do/our-policy-work Young Lives Policy Paper 4: September 2011 Page 3 5. Illness is a recurrent experience for poor households and is associated with other adverse events, such as drought and food shortages. This can result in poorer outcomes for children. For example, in households reporting having experienced environmental shocks, children were half as likely to have a healthy height-for-age in Andhra Pradesh, and households reporting an episode of food shortage in Vietnam and Andhra Pradesh were associated with poorer health in children. 6. Children and caregivers employ a range of strategies to cope with illness, including reducing consumption, selling assets or using savings, labour substitution (by children or others) seeking assistance from relatives or neighbours, borrowing, assistance from NGOs, and use of social protection schemes. Household group membership is associated with better outcomes for children. In households which have no or little access to social support groups, children are 40 per cent less likely to have a healthy weight (Body Mass Index-for-age) (Andhra Pradesh) and healthy height-for-age (Ethiopia and Peru), and children rate themselves lower in subjective well-being (Ethiopia) and health (Andhra Pradesh). 7. Poor families face many barriers in accessing healthcare, including fees for consultation and medicine, distance to the clinic, taking time away from other household activities and concerns over the quality of treatment received. In Ethiopia and Andhra Pradesh families are often heavily dependent on informal sources of social support from family and friends to address some of these costs. Social protection schemes can provide some insurance, giving families more options alongside informal systems, and do not entail the same debt traps. www.younglives.org.uk/what-we-do/our-policy-work Page 4 Poverty, Risk and Families’ Responses 1. Introduction Poverty and structural inequalities shape the life chances of many millions of children worldwide. Chronic poverty is a significant risk factor for poorer life chances, as well as increasing the chances of exposure to other risks, such as droughts, illness, poor nutrition, harmful forms of labour and poor quality, or lack of, services. Across the life course these risks may have different impacts, as well as having cumulative effects lasting into adulthood, reinforcing the transmission of poverty and inequalities to the next generation (Boyden and Cooper 2009; UNICEF 2010; Wachs forthcoming). Understanding the nature of risks faced by children, their households and communities, how they intersect to affect children over the life course and which children are most ‘at risk’ is therefore essential in breaking cycles of poverty and inequalities (Harper et al. 2003). At the same time exposure to risk does not necessarily always result in negative outcomes. Among children growing up in adverse circumstances some fare better than others (Anthony 1987; Luthar et al. 2000; Werner and Smith 2001). These children are understood to be competent or resilient in the context of adversity (Masten et al. 1990; Rutter 1985). Understanding what are the sources of support and protection for children and how these can be strengthened to improve life chances is also fundamental for policies aimed at preventing the intergenerational transmission of poverty. Young Lives has been following two cohorts of children in four countries since 2002 with a view to understanding the causes and consequences of childhood poverty. After three rounds of survey and qualitative data collection Young Lives offers important insight into the interaction of risk and protective processes for children’s life chances and developmental pathways. This paper draws on data from Young Lives, focusing in particular on the trajectories of the older cohort, aged 14–15 years old in 2009, given the increasing focus on adolescents and youth, as illustrated by the World Bank’s 2007 World Development Report and the most recent UNICEF State of the World’s Children 2011. The core argument which runs throughout this paper is that poverty and inequalities are at the heart of childhood risk, shaping which children are at risk, access to sources of protection and therefore children’s life chances. This paper supports the need to shift from averages, as encompassed within the Millennium Development Goals (MDGs), to include measures of equity in order to monitor whether all children benefit from development (Vandemoortele 2011). These findings are timely given current discussions on what may follow the MDGs after 2015 (Dornan 2010a; Kabeer 2010). This paper accompanies a series of detailed country reports which analyse the first three rounds of survey data (Escobal et al. 2011; Galab et al. 2011; Le Thuc et al. 2011 and Woldehanna et al. 2011), and two policy papers (Dornan 2011 and Pells 2011). This paper focuses in-depth on child-level outcomes which occur in the context of the broader trends in consumption growth, wealth and inequality discussed in Policy Paper 5 (Dornan 2011). It begins with an overview of the Young Lives study and definitions of risk, resilience and protection, followed by two key sections. The first section examines how disparities in children’s life chances widen over the life course, with risks being cumulative and compounded, frequently disadvantaging the same children. The second section explores the complex interconnections between risk, protective processes and children’s life chances, through the lens of family illness and death. The paper concludes by considering implications for policy. www.younglives.org.uk/what-we-do/our-policy-work Young Lives Policy Paper 4: September 2011 Page 5 2. About Young Lives Young Lives is a longitudinal study of childhood poverty which has been collecting both survey and qualitative data on around 12,000 children since 2002 in four countries (Ethiopia, the Indian state of Andhra Pradesh, Peru and Vietnam). The study comprises: ●● An Older Cohort of around 1,000 children in each country (slightly fewer in Peru) born in 1994/95. Household and child-level data were collected on this group when they were 8, 12 and 15 years old. In-depth qualitative information was collected on 25 older cohort children in each country in 2007 and 2008 when they were 13 and 14 years old. In Andhra Pradesh the third round of qualitative data collection was completed in 2010 and is drawn upon in this paper. A Younger Cohort of around 2,000 children in each country born in 2001/02. Survey data were collected on this group when they were 2, 5 and 8 years old. Qualitative information was collected on 25 younger cohort children per country in 2007 and 2008. ●● Young Lives has also carried out various qualitative sub-studies, plus research into school quality and access in Ethiopia in 2009 and in Andhra Pradesh in 2010. All the names of Young Lives children used in this report are pseudonyms. The study is pro-poor, meaning it focuses on children growing up in poorer populations.1 The data therefore are not nationally representative, nor should they be compared simplistically between the Young Lives countries, since the samples of children are drawn differently. What such panel data can be used to do is highlight how change over time is affecting children and what similar (or different) processes are going on in each of the Young Lives countries, as well as disparities within the samples. 3. Risk, resilience and protection The concepts of risk and resilience have become increasingly popular in understanding the impact of negative circumstances on children’s development, and factors which enable children to ‘recover from, adapt to and/or remain strong in the face of adversity’ (Boyden and Cooper 2009; UNICEF 2011). While risk is understood as factors which increase the probability of negative outcomes, there is much less conceptual and methodological clarity over resilience. Resilience is broadly understood as being the capacity to overcome negative circumstances, but the ways in which the concept has been used in research and policy have been criticised, in particular the focus on individual internal psychological processes and wellbeing to the neglect of broader structural factors and social processes (Boyden and Cooper 2009; Boyden and Crivello forthcoming). Individualist approaches shaping policies target symptoms rather than the root causes of the problem and ‘de-politicises the project of poverty reduction’ (Boyden and Cooper 2009; Seccombe 2002). This focus contradicts research in developed countries which has demonstrated that socio-economic adversity is one of the principal risk factors for child development (Feinstein 2003; Schoon 2006). Moreover, it tends to result in a focus on categories of children ‘at risk’, such as orphans and vulnerable children, and crisis situations 1 Details can be found in technical notes which describe the sample at http://www.younglives.org.uk/our-publications/technical-notes. www.younglives.org.uk/what-we-do/our-policy-work Page 6 Poverty, Risk and Families’ Responses such as humanitarian disasters, rather than the underlying structural causes of risk (Smart 2003; UNICEF 2011). The subjective perceptions and collective understandings of those affected also tend to be overlooked in favour of ‘objective’ standards of negative and positive developmental pathways (Boyden and Crivello forthcoming). Research from Young Lives has challenged a number of these assumptions regarding what constitutes risk for poor children. Risk is ‘not simply a feature of “extraordinary” childhoods and “extraordinary” circumstances but also an integral part of everyday “ordinary” lives in which the young negotiate multiple, interacting challenges’ (Crivello and Boyden 2011: 1). Challenging assumptions about risk and vulnerability Alongside the main study Young Lives is conducting a sub-study funded by the Oak Foundation looking at risk and vulnerability. In Ethiopia this focuses on orphans and vulnerable children and in Andhra Pradesh on child work. Both studies seek to better link research with policy by questioning prevailing assumptions about what puts children at risk. In Ethiopia (along with sub-Saharan Africa in general) the HIV/AIDS pandemic as well as conflict in many countries has led to a focus on orphanhood and vulnerability. Several countries including Ethiopia have created national-level policies for ‘orphans and vulnerable children’. Yet it is not apparent which are the most ‘vulnerable’ children and whether orphanhood is the main source of vulnerability. Rather than parental death, the study revealed that where children live (urban or rural environment) and the economic status of the household were more likely to predict better outcomes for children in terms of school enrolment and healthy BMI and height (Crivello and Chuta, forthcoming). However, orphans tended to rate themselves lower in subjective well-being, suggesting that parental loss impacts upon children in other ways, such as their identity, and effects may appear later. In India, child labour and children’s work are very contentious issues, especially with campaigns to include agricultural work in the government’s list of hazardous work. The sub-study in Andhra Pradesh focuses on what children view as hazardous and non-hazardous work and on the strategies adopted for coping with risks in daily life. Children reported that injuries at work are outweighed by injuries that occur in leisure time, corporal punishment in schools and actual or feared road accidents (Morrow and Vennam, forthcoming). Moreover, children believe that it is important to work to support the household, and some question whether school will equip them with the skills necessary to earn a living. This suggests that rather than focusing on work as an isolated phenomenon it would be better to consider the environments more broadly in which children and adults are both living and working. Findings from both studies challenge policymakers, especially within the field of child protection, to move beyond a focus on categories of children at risk, such as orphans and child labourers. Instead they encourage policies aimed at addressing the underlying causes of poverty and inequality which put children at risk of poorer outcomes and hazardous situations. In response this paper takes a broad view of risk and resilience, one which encompasses broader structural processes and group functioning, alongside the child and household level. Although risks can have specifics effects for children, given their age and stage of development, risk and protection cannot be understood in isolation from their families and communities and how these are shaped by socio-economic stratification and the unequal distribution of power and resources. Given the theoretical and functional difficulties with the concept of resilience, this paper uses the term protective processes, suggesting ongoing actions which may moderate and mitigate the impact of risk on children and their families. In this context factors may simultaneously represent risk or protection, implying that they can either contribute to or undermine life chances. For example, children may leave school in order to work to support the household because of the death of a parent. This may be a risk factor for the longer-term chances of the young person obtaining a higher-paid job and potentially may be damaging to their health and sense of well-being. At the same time this strategy stops the household from either disintegrating, starving or having to take out a loan at a high www.younglives.org.uk/what-we-do/our-policy-work Young Lives Policy Paper 4: September 2011 Page 7 interest rate, and the child may develop more vocational skills and a sense of achievement at supporting the family. This is part of a series of trade-offs that families living in chronic poverty are forced to make (Kabeer 2000), especially in coping with adverse events, as emerges in the following sections. The paper uses a mix of quantitative and qualitative methods and captures both the broader trends and the more complex and nuanced nature of socio-cultural processes and interaction between risk and protection. It attempts to balance objective indicators of well-being with the subjective understandings of children and their caregivers.2 Recognising the methodological challenges in capturing the ‘sheer multiplicity of potential risk and protective factors and the possible relationships among them’, this paper is cautious in attributing exact causation (Barton 2005: 142). Instead, it attempts to identify channels through which factors may have a negative effect on children’s life chances, such as lower levels of cognitive development or poor health, or which may assist children and families in responding to risk. 4. Disparities in life chances widen over the life course Across the four countries in the Young Lives study there have been falls in absolute poverty (Dornan 2011). However, this has not translated into better life chances for children at the same rate and there are significant disparities between groups of children. This is illustrated in the following sections, which examine the changes in first, school enrolment across the three rounds of survey data collection and second, cognitive achievement between 2006 and 2009 of children in the Older Cohort. Globally, huge strides have been made towards reaching MDG 2, which focuses on universal free and compulsory primary education. Within Young Lives study countries primary schooling has increased, particularly in areas and among groups with initially lower enrolment levels (Pells 2010). Yet rising enrolment rates do not fully capture children’s pathways through schooling and how and why these may vary between different children, understanding of which is critical to policy improvements in education. Across all four countries household wealth (a composite measure which includes consumer durables, quality of housing and access to services) is a predictor of school enrolment. As the children become older those from the poorest households are least likely to be enrolled and have lower levels of cognitive achievement. For example, in Vietnam the percentage point difference in school enrolment level between poor children and the rest more than tripled between the ages of 12 and 15. Poor children are also more likely to have to repeat a grade at school, probably because of periods of prolonged absence and poorer performance as a result of the poorer quality of teaching. In Peru, 61 per cent of the Older Cohort in the bottom quintile of household consumption have had to repeat a grade by 2009, while in the top quintile 38 per cent of children have repeated a grade. With the exception of Vietnam (where the largest single difference is between children from ethnic majority and minority groups) the largest disparity in school enrolment in 2009 was between children in urban and rural areas. In Andhra Pradesh (Figure 1) and Vietnam this gap 2 The term ‘caregiver’ has been used in recognition of the fact that children are not always cared for by their biological parents. Where a quotation has been attributed to a caregiver this individual is either a relative or guardian but not the biological parent of the child. www.younglives.org.uk/what-we-do/our-policy-work Page 8 Poverty, Risk and Families’ Responses appears in 2006, only to have increased by 2009, whereas in Peru an urban/rural difference does not appear until 2009. In Ethiopia, school enrolment in rural areas was lower than in urban areas in 2002 (as many children were not yet in school at age 8) and widened with each subsequent round of data collection. The different patterning between the countries may reflect the different stages of economic development and social policy, for example, Ethiopia having weaker service infrastructure hence having the highest percentage of rural children who never attended school. Figure 1. Widening gaps in school enrolment, Andhra Pradesh Older Cohort, 2002-9 (n. = 1008) 100 Rural Percentage of children enrolled in school 95 Urban 90 85 80 75 70 2002 2006 2009 Multiple factors intersect in causing children to discontinue their schooling. The following story of Mohan illustrates how poverty and shocks can result in children having to leave school in order to work to support their family. In rural areas the need tends to be greater, not only because of higher poverty rates and a greater prevalence of economic and environmental shocks, but also because household livelihoods depend on labour-intensive agriculture. Yet Mohan’s case also raises the question of the appropriateness of education and whether in the longer term it enables children in his situation to earn a living. This underscores the need to consider education in the broader context of economic development and whether it is equipping children with the skills that they need in order to enter the local labour market. Learning is not always only about schooling Mohan is 15 years old and from a Backward Caste in rural Andhra Pradesh. He stopped going to school after the seventh grade as his father was sick. He did not return to school, even though his parents encouraged him to do so. He said ‘everyone will be new and I thought it will be awkward if I go back so I never went.’ Previously he used to be beaten by teachers as he was absent frequently. Mohan also explains that in the future he will be responsible for supporting his parents as well as his own family when he marries and says staying in school is ‘risky’. He believes it is important to learn skills like his father, who is a mason, so that he can earn money to have a good life. Mohan lists the range of skills he has learnt from his father and older cousins including ploughing, driving a bulldozer, driving a bull cart, mending the tractor, building walls and plastering. www.younglives.org.uk/what-we-do/our-policy-work Young Lives Policy Paper 4: September 2011 Page 9 While important, differentials between girls and boys are much smaller than between children living in urban areas and those living in rural areas (Pells 2011; Dercon and Singh forthcoming). This is a situation reflected across the countries. However, important trends are emerging as the children become older which have implications for the reproduction of inequalities and children’s transition beyond school to livelihoods, the labour market and the establishment of households of their own. Though within the Young Lives sample there was parity in the numbers of boys and girls going to primary school, national data show that though this has been achieved in Vietnam and Peru, gender parity does not yet exist at primary level in Ethiopia and India (UNICEF 2010). By 2009 differences between boys and girls have become clearer yet, contrary to the dominant narrative of girls being consistently disadvantaged (Levine et al. 2008; Jones et al. 2010), within the Young Lives sample it is not always girls who fare less well. In Peru, Vietnam and Ethiopia more boys than girls in the Older Cohort are leaving school by the age of 15 years. This could be due to the higher wage-earning potential of boys, or boys performing less well in exams (UNESCO 2010: 67). In addition girls may be able to combine tasks at home with studying, but boys may find it harder to combine school with work, given the different types of work undertaken according to gender. (This is explored in much greater depth by Pells 2011.) Contrary to reports of caregivers with low levels of education having low aspirations for their children, and to ‘poverty of aspirations’ debates (see, for example, Appadurai 2004), across the countries in the Young Lives study caregivers frequently expressed positive attitudes towards children’s education. Women in particular were keen for daughters to have opportunities denied or unavailable to them when younger: In the past girls married at the age of 15 but now things have changed. Girls can marry at the age of 20 and even 30 years. My wish about my daughter is that she should marry after she has become self-reliant; I wish her to complete her education, then to have her own work and then to marry a person whom she loves and with whom she wants to live. (Caregiver [grandmother], aged 70, rural Ethiopia) Fathers and brothers were also positive about their daughters and sisters studying, as illustrated by the following interview conducted in rural Andhra Pradesh. Interviewer: In your village, as soon as the girls grow up, they get them married. Will you do like that? Father: No I will not do like that. If she [daughter] wants to study, we will not stop her but if she stops studying then I will get her married. If she says she wants to study further, I will spend the money and will make sure she fulfils her wish, if she says she wants to stop studying then I will get her married, that is it. At the same time, caregivers are concerned about the likelihood of children being able to get better-paid jobs, even after completing education. This leads some parents to question the value of keeping children in school: ‘Even if she were educated it is still not possible to get a job… Then what’s the point in getting schooled, no schooling can get her a job’ (mother, rural Andhra Pradesh). In this case both the son and daughter had left school in order to assist the mother with earning enough to support the household. There are also concerns specific to daughters concerning future marriage prospects: If she were educated I mean even if she were to have passed the tenth grade people would be apprehensive [in making an offer of marriage]… they would say that educated girls don’t know how to work manually. With this excuse they would demand maybe 50,000 rupees as a dowry. From where can I get such a huge amount of money? (Mother, rural Andhra Pradesh) www.younglives.org.uk/what-we-do/our-policy-work Page 10 Poverty, Risk and Families’ Responses Processes of social change are taking place, bringing opportunities, particularly in education; with children able to study longer than their parents (Crivello 2011). At the same time, being on the cusp of a wave of social change entails a degree of social risk for families whose livelihoods and social reputation may be at stake if they are out of step with the wider community (Boyden and Crivello forthcoming; Boyden, Pankhurst and Tafere forthcoming). 4.1 Risks can be cumulative and compounded for poor children Risks factors can be cumulative and compounded, interacting with one another and thereby having an increasingly negative impact on children’s life chances over time. This is illustrated in the following section, which examines children’s outcomes in formal learning. Children’s experiences of school and formal learning outcomes are shaped not only by their socioeconomic status and household situation, but also by factors within school. These can include the quality of education received, challenges in the school environment such as lack of gender-segregated toilets or forms of corporal punishment, as well as differing attitudes and expectations of teachers for boys and girls (Pells 2011; Rojas 2011; Streuli et al. 2011). Factors internal and external to the school environment may therefore hinder progression through school, increasing the risk of grade repetition and of children leaving school before they complete a full course of primary education. This section therefore explores the compounded nature of risk through the intersection of poverty with factors within the school environment. Controlling for other factors, the household position on the wealth index is associated with children’s cognitive achievement. Cognitive achievement is measured for all children, not just those in school, and the score used here combines two measures, one testing receptive vocabulary3 and the other testing maths. The wealth index is constructed on a scale with 0 (minimum) and 1 (maximum). Across all four countries children who live in wealthier households performed better in the cognitive achievement tests. Each point on the wealth index (i.e. moving from the bottom to the top) accounted for 1.6 points on a five-point cognitive score of children in Andhra Pradesh, 1.7 points in Peru, 2 points in Vietnam and 2.2 points in Ethiopia. The following graph should be used to look at the patterns within each country rather than to compare literacy levels between countries given sampling differences. The graph shows that literacy rates, defined as the ability to read a single sentence without difficulty, rises with household wealth level, so that children in the poorest households have the lowest literacy rates. In both Vietnam and Peru there is a large gap between children in the poorest quintile and the rest. Further analysis suggests that this may be due to the marginalised status of ethnic minorities within the countries. This is illustrated by Figure 3, which shows that in Vietnam children from ethnic minorities are over three times more likely to be in the lowest category of cognitive achievement than their Kinh peers (about 62 per cent of children from minority ethnic backgrounds were seen as having low cognitive achievement on the measure shown in the chart; 3.4 times the percentage of the Kinh children in this group in 2006). Moreover, while the numbers of Kinh children in the bottom category declines between 2006 and 2009, there is no change among children from ethnic minorities.4 3 Receptive vocabulary (school readiness), verbal skills and achievement were measured using tests Young Lives developed or adapted from standardised international tests, such as the Peabody Picture Vocabulary Test (PPVT). We acknowledge that bias may arise when testing children with different languages and cultures using the same instruments, although measures were taken to adapt them to local contexts and languages and in no case were original standard scores used. Bias is an especially important consideration in testing children who speak minority languages. Reliability and validity results for our test administrations and concerns are presented and discussed in Cueto et al. (2009a). In particular, the authors of this document recommend that results should not be compared across countries, or across groups with different maternal languages within countries. 4 The picture for children from ethnic minorities could be even bleaker. When wealth is controlled for, Kinh children are 80 per cent less likely than children from ethnic minorities to be enrolled in school. www.younglives.org.uk/what-we-do/our-policy-work Young Lives Policy Paper 4: September 2011 Page 11 Figure 2. Increasing literacy levels by wealth quintile (12-year-olds, 2006 %) Poorest 100 Percentage able to read a single sentence without difficulty 90 80 70 60 50 40 30 20 10 0 Ethiopia India Second Middle Fourth Richest Vietnam Peru Figure 3. Stark disparities in percentage of Older Cohort children in the lowest and highest categories of cognitive achievement in Vietnam, according to ethnicity, 2006 (n. = 925) and 2009 (n. = 961) Kinh lowest 70 60 50 40 30 20 10 0 Kinh highest Minority Ethnic lowest Minority Ethnic highest Percentage of children 2006 2009 Note: Cognitive achievement is derived by combining performance on mathematics and PPVT scores to produce a ranked variable which identified broad cognitive achievement.5 5 This is intended to give a stylised understanding of cognitive achievement, using the available data. To produce this estimate, first PPVT and maths scores were separately coded into low, medium and high scores. The composite was then produced by combining these variables (low PPVT and low maths = lowest category; low PPVT and medium maths or medium PPVT and low maths = mid-low; low PPVT and high maths or medium PPVT and medium maths or high PPVT and low maths = middle; medium PPVT and high maths or high PPVT and medium maths = mid-high; high PPVT and high maths = highest). www.younglives.org.uk/what-we-do/our-policy-work Page 12 Poverty, Risk and Families’ Responses The poor cognitive achievement of children from ethnic minorities may be attributed to the higher level of drop-out among children from ethnic minorities, but also to the challenges of effectively implementing bilingual education and the socio-economic pressures faced by this marginalised community. A similar situation exists in Peru, where only 10 per cent of children from ethnic minorities attend bilingual schools, the rest receiving education in Spanish. Young Lives research also found that half the teachers in bilingual schools in southern Peru could not speak the local indigenous language and that bilingual materials were not being used (Cueto et al. 2009b). These disparities are indicative of an ‘inverse care law’ whereby poorer and more marginalised children often attend schools with fewer resources and receive a poorerquality education (UNESCO 2010: 23). While there is little movement between the categories of cognitive achievement between 2006 and 2009 in the graph above, in Ethiopia the situation is more fluid. Figure 4 demonstrates how the number of children in the lower and higher categories of cognitive achievement rises between 2006 and 2009, suggestive of slightly increasing polarity. 6 Figure 4. Increasing polarity in cognitive achievement scores between 2006 and 2009, Ethiopia Older Cohort 250 2006 2009 200 Number of Children 150 100 50 0 Lowest Mid-Low Average Mid-High Highest Table 1 explores this polarity in cognitive achievement scores further by showing that a greater proportion of urban children and boys moved into the highest category between 2006 and 2009. In comparison, more children in rural areas and girls moved into the lowest category. 6 A higher percentage of children in the Older Cohort completed both tests in 2009 (n. = 671), which accounts for the higher number of children recorded on the graph in 2009 compared to 2006 (n. = 665). www.younglives.org.uk/what-we-do/our-policy-work Young Lives Policy Paper 4: September 2011 Page 13 Table 1. Increasing percentages of girls and rural children performing less well between 2006 and 2009 (Ethiopia, Older Cohort) Percentage of children in lowest and highest categories of cognitive achievement 2006 (n. = 665) Entire sample Boys Girls Rural Urban Lowest Highest Lowest Highest Lowest Highest Lowest Highest Lowest Highest 17 16.6 16.2 17.9 17.6 15.4 24.7 7.3 6.6 29.1 2009 (n. = 671) Lowest Highest Lowest Highest Lowest Highest Lowest Highest Lowest Highest 19.4 17.4 17.3 20.2 21.6 14.5 28.1 7.7 7.2 31 In Peru and Ethiopia boys are increasingly outperforming girls in cognitive achievement tests over time. Conversely Vietnam reveals a different picture, with the number of Young Lives girls in the highest quintile of performance increasing, while the number of boys is decreasing. In Andhra Pradesh there is a consistent pattern of girls being disadvantaged, with boys 1.8 times more likely to be enrolled by 2009 and performing better in cognitive achievement tests. This may become increasingly significant in the future, given the importance of maternal education for Young Lives children. Studies have shown maternal education to be a significant factor in breaking the intergenerational transmission of poverty, as girls who stay in school for longer marry and have children later, so decreasing the risk of maternal and child mortality and morbidity (Jones et al. 2010; Levine et al. 2008; Lloyd with Young 2009; Temin and Levine 2009; UNICEF 2010: 4). Research has demonstrated that children’s school attainment is also associated with the education level reached by their mothers (Levine et al. 2008: 20). Children with mothers with no or little education tend to have lower levels of cognitive achievement, progress through school more slowly and leave school earlier. This is reflected by children in the Older Cohort in Peru: ●● Three times the number of children whose mothers completed primary school left school between 2006 and 2009 than children whose mother received no education. Two-thirds of children whose mother had no education had repeated a grade by 2009, in comparison with two-fifths of children with mothers who at least completed primary school. When considering children with mothers who progressed to higher education, none had left school by 2009 and only 14 per cent repeated a grade. ●● ●● More analysis is required to understand the relationship between maternal education and household socio-economic status. Although maternal education is a fairly robust proxy for other household factors, it is likely that it reflects rather than wholly drives household disadvantage. In other words, households with a history of disadvantage also tend to have mothers with lower levels of education. However, using regression analysis which controlled for a series of variables (see section 5) including household wealth, higher maternal education still found an independent effect attributed to consumption or wealth, associated with higher www.younglives.org.uk/what-we-do/our-policy-work Page 14 Poverty, Risk and Families’ Responses child cognitive achievement across all four study countries. Each year of maternal education accounted for increases on the cognitive achievement scale (measuring between 1 and 5) of 0.02 in Ethiopia, 0.06 in Vietnam, 0.07 in Andhra Pradesh and 0.08 in Peru. This may be because more highly educated mothers are better able to support children’s formal learning and have higher bargaining power for resources within the household supporting educational and healthy development (Levine et al. 2008: 20). In summary, Young Lives data on children’s trajectories through school and outcomes in formal learning demonstrate that disparities increase over children’s lives between non-poor children and poor children, between children living in urban and rural areas, and between groups with different social status because of ethnicity or caste. The differences between boys and girls also increase over time, although at a much lower rate than the other disparities. However, maternal education appears as a significant mediating factor, as children with more highlyeducated mothers have better life chances. Risks are cumulative and compounded across the life course but often in more nuanced ways than commonly assumed, as illustrated by the complex picture concerning gender differences. 4.2 Chronic poverty and inequalities repeatedly disadvantage the same children Chronic poverty and inequalities repeatedly disadvantage the same children, as growing up in poverty can have long-term implications for children’s physical and cognitive development and in turn for their life chances. This section brings together the education-related areas of the previous two sections with indicators of healthy development and nutrition, to illustrate how the same children are at a higher risk of having poorer outcomes in multiple areas of education, health and subjective well-being, due to poverty and inequalities. Inadequate nutrition, including insufficient food, poor micronutrient intake and disease at an early age can result in stunted growth (Dornan 2010a). Young Lives data have demonstrated a number of consequences of stunting including poorer cognitive development in Vietnam (Le Thuc 2009) and Peru (Sanchez 2009). Stunting also impacts on children’s subjective wellbeing, with children reporting higher levels of shame or embarrassment because of being shorter than their peers, and so is likely to impact on children’s progression through school (Dercon 2008). It can also increase susceptibility to physical illness, as the body is less able to fight infections. Stunting therefore has long-term implications for children and also for the wider society, through lost potential. As with inequalities relating to education, certain groups of children are more likely to be stunted than others. In all countries except India (Andhra Pradesh), higher household wealth is associated with greater chance of having a healthy height-for-age. 7 Similarly, in all four Young Lives study countries children in urban areas are more likely to have a healthy height-for-age than their rural counterparts (Figure 5). Although the samples are not nationally representative because of the pro-poor sampling and therefore findings are not directly comparable between countries, it is important to observe the similar levels across the countries, despite very different stages of economic development – Peru has four times the GDP per capita of Vietnam (Dornan 2011). Economic growth, while important, does not necessarily translate into better outcomes for children at the same rate. For example, in Andhra Pradesh, while GDP doubled between 2002 and 2009, the stunting rate only fell by four percentage points (Dornan 2011). 7 BMI-for-age and height-for-age were calculated using international norms established by the World Health Organization. Children classified as having a healthy BMI and height fall within two standard deviations of the norm. www.younglives.org.uk/what-we-do/our-policy-work Young Lives Policy Paper 4: September 2011 Page 15 Figure 5. Percentage of children having healthy height-for-age, by location, 2009 100 Urban 90 Percentage having healthy height-for-age 80 70 60 50 40 30 20 10 0 Andhra Pradesh Ethiopia Peru Vietnam Rural The chart suggests that processes of economic growth and development have been unequal, with children in rural areas having lower access to (quality) services. Ethnicity is also a factor, as there is a higher prevalence of stunting in Peru, Vietnam and Andhra Pradesh among children from ethnic minorities or lower-caste status, when controlling for other factors. Children who belong to the Scheduled Caste in Andhra Pradesh are 70 per cent less likely to have a healthy height-for-age, whereas in Vietnam children are 3.3 times more likely to be in the healthy range if they belong to the Kinh (majority) ethnic group. While groups with low social status tend to live in rural areas, the stunting rates for these groups are higher than the average for rural areas. This may be due to the difficulties associated with providing quality services in remote areas. It may also be a result of political economy whereby marginalised groups have less voice in demanding resources. Objective indicators of healthy development are associated with subjective indicators. Children were asked to rate their health as better, the same, or worse, than others. Across all four countries, controlling for other factors, children who reported worse health than others were more likely to be stunted. Better health is also associated with better outcomes in other areas. In Andhra Pradesh and Vietnam children who rated their health as better than others were more likely to be enrolled at school and have higher cognitive achievement scores. In Ethiopia higher self-rated health was also associated with more chances of being enrolled. Poverty is also associated with lower subjective well-being. Young Lives uses a ladder exercise to collect data on children’s perceptions of well-being. Children are asked to position themselves on a ladder where the ninth step represents the best possible life and the first step represents the worst. Across all four countries, controlling for other factors, higher household consumption levels are associated with children positioning themselves higher on the ladder. This suggests that poverty and inequalities impact on children’s subjective well-being. Notably, higher levels of maternal education are associated with a higher position on the ladder in all countries, apart from Vietnam, and higher self-rated health in Andhra Pradesh, Peru and Vietnam.8 8 Research based on Young Lives data from Peru found that children with lower subjective well-being, in this case perceptions of being poorly respected, had lower cognitive achievement (Outes et al. 2010). This reiterates the link between objective and subjective indicators of child well-being. www.younglives.org.uk/what-we-do/our-policy-work Page 16 Poverty, Risk and Families’ Responses The overlap between better outcomes in enrolment and formal learning and indicators of health is illustrated more clearly by moving from looking at averages for groups of children to individual children. Table 2 disaggregates the number of cases of children in Peru with poor outcomes in one of the education and formal learning domains (enrolment or low cognitive achievement) and in proxies for health (out of the healthy range for BMI-for-age or height-forage). The key issue is whether children from particular groups are over-represented. Stark inequalities are highlighted by the over-representation of children from rural areas, ethnic minorities and households in the poorest quintile of household consumption. The make-up of the group of children showing less good enrolment, cognition and health indicators compared with the whole sample suggests that children from rural areas are 2.5 times less likely to fare well in life chances in both education and health. Similarly, children from ethnic minorities (1.5 times) and children from households in the poorest quintile of household consumption (1.3 times) are less likely to fare well. A similar picture emerges in Vietnam with children from ethnic minorities (over 1.5 times) and from households in the poorest quintile (1.3 times) being less likely to fare well in education and health. This suggests that in both countries there are highly marginalised groups, as discussed earlier with regards to literacy (Figure 2). Table 2. Over-representation of children from rural, ethnic minorities and poor households with poor outcomes in both education and health, Older Cohort, Peru, 2009 Children not enrolled or poor cognitive achievement AND out of healthy range of height-for-age or BMI-for-age Group Boys 53.6% of sample Girls 46.4% of sample Rural 23.5% of sample Urban 76.5% of sample Mestizo 78.8% of sample Ethnic minorities 17.3% of sample White/Asian 4% of sample Consumption quintile 1 (lowest) Consumption quintile 2 Consumption quintile 3 Consumption quintile 4 Consumption quintile 5 (highest) % 52.6 47.4 59.4 40.6 70.6 25.4 4.0 26.4 21.2 20.8 17.9 13.7 Number of cases 162 146 183 125 214 77 12 81 65 64 55 42 Note: This table covers two-fifths of the sample (308 children). In terms of scale, children with poor outcomes in one health-related and one education-related dimension are much greater in number in Andhra Pradesh and Ethiopia, with just over twothirds of children in the Older Cohort in Andhra Pradesh and just under two-thirds of children in Ethiopia falling into this category. However, children from the poorest quintile of household consumption are only slightly over-represented in Andhra Pradesh and Ethiopia and not to the same extent as in Peru and Vietnam. In Andhra Pradesh children from urban areas are a quarter less likely to have combined poor outcomes, and children not from the low-status caste groups are over a fifth less likely to have combined poor outcomes. In Ethiopia, the largest disparity is between children from urban and rural areas, with children from rural areas being 1.2 times more likely to have combined poor outcomes. www.younglives.org.uk/what-we-do/our-policy-work Young Lives Policy Paper 4: September 2011 Page 17 This may reflect patterns of inequality within these countries. While poverty rates have declined within all four countries, in India and Vietnam inequality has increased. In Peru inequality increased between 1990 and 2000 but declined in the period thereafter, until 2008. In Ethiopia inequality fell (Dornan 2011). This reiterates the need to ensure that economic growth results in an improvement of material circumstances and a reduction in cumulative and compounded risk for the poorest and most marginalised groups in society. Key messages ●● Disparities in life chances widen over the life course. Despite the different stages of economic development and socio-cultural contexts of the four countries, there is a consistent pattern of disparities, which widen over time. Household wealth is a significant predictor of continued school enrolment and cognitive achievement. As the children become older, differences in school enrolment and cognitive achievement between non-poor children and poor children, between children living in urban and rural areas, and between groups with different social status because of ethnicity or caste, increase over the course of children’s lives. Risks are cumulative and compounded. Enrolment figures, for example, mask absence from school because of illness, caring for others or work for the household. This may hinder progression through school, and be compounded by challenges in the school environment, with both increasing the risk of grade repetition and children leaving school before they complete a full course of primary education. For children in school there is an ‘inverse care law’ where the poorest and most marginalised children receive a lower quality of education (whether because of the poor quality of the school environment, being taught in a language other than their mother tongue, or inflexibility in the school system so that children are unable to balance the competing demand on their time from home and school). Consequently, in order for children’s and parents’ high aspirations to be reached, it is important that school equips children with the skills needed to access the labour market, vocational training and higher education. Process indicators such as enrolment should be accompanied by indicators of quality, as contained within the Education For All goals to promote the learning and life skills of young people. Gender differences are often smaller and more nuanced than other disparities. Andhra Pradesh has a more consistent pattern of girls being disadvantaged, but there are different patterns around enrolment and cognitive achievement in the other three countries (Dercon and Singh forthcoming). This mirrors the global picture where it is observed that ‘gender gaps in education and health are small compared to gaps in outcomes across ethnicity, socio-economic class or geographical location’ (World Bank 2011: 5). However, the association between higher maternal education and better outcomes for children in Young Lives data suggests that gender differences may become more significant later in the life course. Chronic poverty and inequalities repeatedly disadvantage the same children. Children from ethnic minorities/low-caste groups and from rural and the poorest households are overrepresented among children who have poorer life chances in education and health. This suggests an increased risk of doing less well in multiple dimensions for children from these backgrounds, and also suggests the way in which risk is mediated through poverty and structural disadvantage. ●● ●● ●● www.younglives.org.uk/what-we-do/our-policy-work Page 18 Poverty, Risk and Families’ Responses 5. Poverty and inequalities are at the root of risks for children The nature of chronic poverty is such that it is enmeshed with multiple risks. Children and families living in chronic poverty have fewer resources to cope with risks or adverse events such as illness and death, meaning that they are hit harder, which in turn can perpetuate the cycle of poverty and risk. This can have specific and long-lasting consequences for children’s life chances because of their age and stage of development. At the same time, where possible households draw on a range of strategies to cope with and respond to risk. Children play a key role in managing risk within the household (Vennam et al. 2010). Yet access to these protective processes is frequently determined by the same inequalities of power and resources (such as income, livelihoods, assets, credit, insurance, quality of the living environment and access to quality services) that give rise to higher risk for poor families. The following sections explore the interconnections between risk, protective processes and children’s life chances, through the lens of family illness and death, selected because of their prevalence in the everyday lives of poor children and families. Regression analysis is used to examine the impact of a series of potential risk and protective factors reported in 2006 upon enrolment, cognitive achievement, self-rated health, BMI-for-age, height-for-age and subjective well-being measured in 2009 of children in the Older Cohorts across all four countries. This includes the socio-demographic characteristics of the child and household, risk factors (including economic and environmental shocks, illness and death of a family member) and protective processes (including household membership of groups and level of support, and access to services). Although this analysis does not necessarily prove causality it can be used to illustrate factors which are associated with one another, supported further by interviews with children, their caregivers and community members, from Ethiopia and Andhra Pradesh. 5.1 Poor households are affected by multiple and recurrent experiences of illness Poor households are affected by multiple and recurrent experiences of illness. Table 3 illustrates the prevalence of illness reported by households across the four countries. It is likely that there is under-reporting of illness because of it being a common part of daily life, especially among the poorest households, and because of inability to pay the direct or indirect costs (Sauerborn et al. 1996b). The table highlights the greater prevalence of illness shocks among the poorest households, especially in Ethiopia and Andhra Pradesh. Reporting of deaths is also highest in poor households, for example, in Andhra Pradesh 9.4 per cent of poor households report a death, compared with 7.2 per cent of non-poor households. 9 9 It should be borne in mind that, given the pro-poor nature of the sampling, even those households defined as ‘non-poor’ (above 50 per cent of the median household consumption level within Young Lives sample, calculated for each country) do not necessarily have high levels of consumption. Consumption level is measured by a questionnaire administered at the household level which has a series of questions to capture the amount spent by the household in various domains, as well as goods produced and consumed by the household. www.younglives.org.uk/what-we-do/our-policy-work Young Lives Policy Paper 4: September 2011 Page 19 Table 3. Percentage of households reporting illness between 2006 and 2009, by country, location and household consumption level Andhra Pradesh All Rural Urban Lowest consumption quintile (2009) Highest consumption quintile (2009) 31.3 33.7 24.3 39.4 28.0 Ethiopia 44.1 44.6 45.2 51.0 44.8 Peru 20.7 19.5 21.6 19.3 19.3 Vietnam 29.0 28.4 32.8 32.0 28.9 In part the pervasive nature of illness can be attributed to the poorer quality of the living environment in which poor children live. Poor-quality living and school environments impact on children’s health Afework lives in Addis Ababa, Ethiopia. He says he likes his local area but is ‘sick of the garbage in our surrounding’ and is ‘afraid of the dirty place that would give us cold’. Similarly Afework describes how ‘there are certain places at school that I avoid. This is the dirty place at school because it smells bad and exposes us to flu.’ Previously he caught typhoid from drinking undistilled water. He believes that having good sanitation is essential to having a good life, along with respect for family and playing football. However, ill-health tends to co-occur with other types of adverse events (often called ‘shocks’) such as food price increases, crop failure or droughts, which also disproportionately affect poor households. In Ethiopia 52.6 per cent of households reporting an environmental shock also reported experiencing a family illness or death, and 53.9 per cent of households reporting an economic shock in 2009 also reported experiencing a family illness or death (Dornan 2010b: 11). Ayu, aged 12 and from a rural area in Ethiopia, described how the price of a quintal of maize had increased from 150 to 600 birr. While her family were able to cope by eating the grain harvested from family land, other families without land could not afford to buy maize. People in these families fell sick, including from malaria and typhoid. During the period 2006– 9 many households reported increases in prices (Dornan 2010b: 9). Households with a child in the Younger Cohort reporting food price increases numbered one in three in Vietnam (32.2%), nearly four in five in Andhra Pradesh (77.3%) and nearly nine in ten in Ethiopia (88.3%) The co-occurrence of economic shocks and illness is illustrated by the following example of Kassaye’s family. In addition, the importance of the timing of illness in communities dependent on agriculture is demonstrated. This is particularly pertinent given the 2011 famine in the Horn of Africa. The significance of seasonality for illness is reflected in findings from Burkina Faso which showed that during the rainy seasons when crops are sown and harvested individuals were less likely to report illness, despite the higher prevalence of diseases such as malaria. In addition, illness was more likely to be treated at home to avoid the financial and time costs of going to a clinic (Sauerborn et al. 1996b). Seasonality increases impact of illness on a household Kassaye is 14 years old and lives in rural Amhara, Ethiopia. His father broke his leg and did not work for six months. As it was the sowing season the household was short of labour and the harvest failed, with the result that the family were forced to go hungry. His mother describes how ‘we were eating 50 kilos of any kind of cereal for three weeks. Now we have reduced and we consume the same amount of cereal for one month.’ www.younglives.org.uk/what-we-do/our-policy-work Page 20 Poverty, Risk and Families’ Responses The family were unable to sell households assets because of the drop in price of sheep and other livestock. As a consequence Kassaye explains that ‘food shortage means I go to school hungry. I cannot properly attend class if I am hungry, so it affects my schooling.’ The links between environmental shocks and illness are likely to become stronger in the face of climate change (World Health Organization 2009). A recent Climate Change Vulnerability Index places three of Young Lives study countries in the top 15: Ethiopia (11), India (2) and Vietnam (13) (Maplecroft 2011). Children and caregivers in Andhra Pradesh and Ethiopia reflected on the changing climatic patterns and the frequency of extreme weather conditions, including drought and floods. Controlling for other factors, environmental shocks are associated with a number of poorer objective and subjective outcomes for children, which may be linked to poor health through food shortages (as in the case of Kassaye) or an increased need for children to work to support the household. Children in households reporting environmental shocks: ●● ●● in Andhra Pradesh were only half as likely to have a healthy height-for-age; in Ethiopia and Peru scored quarter of a point lower in the composite cognitive achievement score. In households reporting an episode of food shortage, children: ●● ●● ●● ●● were 60 per cent less likely to have a healthy BMI-for-age in Peru; scored lower in cognitive achievement tests in Andhra Pradesh and Ethiopia; reported lower self-rated health in Vietnam and Andhra Pradesh; positioned themselves lower on the subjective well-being ladder in Ethiopia and Peru. This demonstrates that shocks and adverse events, whether economic, environmental or health related, are a common experience for chronically poor households. Shocks can be both a cause and a consequence of poverty. Households affected by these events have an increased chance of either remaining or becoming poor (Dornan 2010b). 5.2 Illness is a cause and consequence of poverty Poor households are disproportionately affected by illness and ‘can have particular difficulties accessing healthcare, and… when they seek care, they spend a greater proportion of their… income on treatment than richer households’ (Chuma et al. 2007: 674). As one caregiver in Andhra Pradesh explains: In a month if anyone is affected [in this case by malaria and typhoid], it passes to all the others in the family. When the first one is being treated with medical care and everything and he is almost back to normal, another child becomes affected by the same disease, and then my wife, it is like that. The money we got as income is all spent on medical care. When poor families are hit by illness the household can be pushed further into poverty through the direct costs of illness (user fees at healthcare facilities, cost of medicines and transport to the facility) and indirect costs (lost income and time costs). Chuma et al. describe how if 10 per cent of household expenditure is spent on healthcare the impact is ‘catastrophic’ (ibid.: 673). It is hard to calculate the exact value of indirect costs, but other studies have calculated that it can be 2–3.6 times greater than the direct costs (McIntyre et al. 2006; Sauerborn et al. 1996a). For policymakers this poses a considerable challenge as it is also more difficult to intervene to www.younglives.org.uk/what-we-do/our-policy-work Young Lives Policy Paper 4: September 2011 Page 21 alleviate indirect, as opposed to direct, costs. The impact of direct and indirect costs on families is illustrated by the situation of Harika’s family. The indirect and direct costs of illness impact on the entire household Harika lives in a rural community in Andhra Pradesh with her parents and two brothers. Her family belongs to a Backward Caste. Harika’s eldest brother was recently unwell. One night, Harika’s parents came home to find their son with a high fever. He was taken to a doctor in the town of Raichur and initially diagnosed with malaria. When his condition did not improve, he was admitted to a private hospital in Hyderabad and, later, Bangalore. After a series of tests, it was discovered that there was ‘no circulation of blood on one of the nerves on [his] neck’. The entire ordeal lasted for one month, during which time Harika’s mother stayed by her son’s side and was unable to work the family’s fields. According to Harika: ‘We did not take care of the cotton fields and that was the reason the crop was not good.’ The total cost of the treatment was 150,000 rupees, which the family paid by obtaining a loan from Raichur at a low interest rate. Harika’s maternal cousin helped to transport the family to Hyderabad and provided accommodation during their stay. On another occasion, Harika’s father was unwell and Harika had to miss school in order to help her mother tend the fields. Harika feels that her household is poorer as a result of these events. Family illness and death can have long-term impacts on children. For example, controlling for other factors, children in Andhra Pradesh who lost a family member (other than a caregiver) are 70 per cent less likely to be enrolled in school. In both Ethiopia and Vietnam the death of a father is associated with lower cognitive achievement, perhaps suggesting periods of absence or permanent drop-out from school. Other research from Young Lives found that by the time Ethiopian Older Cohort children were aged 12, one in five had lost one or both parents (Himaz 2009). Children who had lost their mothers were 20 per cent less likely to be enrolled and 27 per cent more likely to be unable to read a complete sentence. In comparison, children who had lost their father rated their subjective well-being lower. If a caregiver or another sibling falls ill or dies, then children often have to take over their roles and responsibilities, either temporarily or permanently, as illustrated by the following case of Rahmatulla. Family illness changes children’s roles and responsibilities Rahmatulla is 15 years old and lives in Hyderabad. His elder brother had been the family’s primary earner until his death just over a year ago. Rahmatulla describes how ‘because of that, when I finished ninth class there were financial difficulties here [at home] and I had to work even in the holidays, so I started working in the shop. I had to take care of the house, my sister is also there.’ Rahmatulla works in a clothes shop to pay for his education and his mother has taken on more embroidery work. The financial situation is made worse by the fact that Rahmatulla’s father has an illness which affects his mobility and Rahmatulla has to buy medication for him every day, at a cost of 50 rupees a time. Less explored in other studies and clearly emerging from Young Lives qualitative data is the psychological burden of illness, actual and/or feared. In Ethiopia, Mulu’s mother was unwell for two months during the last year. While Mulu suffers from headaches, fainting and colds she has not told anyone about this despite it affecting her education, as ‘I do not like to tell anybody’. Seemingly Mulu is concerned about causing further worry to her family. While adults often take illness for granted because of its all-pervasive nature, children list illness as one of their majority concerns in life (Boyden 2009). www.younglives.org.uk/what-we-do/our-policy-work Page 22 Poverty, Risk and Families’ Responses However, both children and caregivers in Ethiopia and Andhra Pradesh express fears about being unable to fulfill their responsibilities within the household, continuing to work and/or study, incurring high costs and dying, all as a consequence of illness. Naomi is 12 years old and lives in rural Ethiopia. She suffers from recurring malaria and worries about the state of her health: ‘Sometimes the illness [malaria] is strong and I am absent from school on these days. This disappoints me. I have a strong fear that the illness may continue deteriorating rather than improving and this may hamper my education and other aspects of my life.’ These fears and anxieties are illustrated further by the situation facing Netsa’s family. Illness as a considerable source of anxiety for children and caregivers Netsa lives in an urban area in Ethiopia with her brother and single mother. Her mother is currently undergoing anti-retroviral therapy and finds herself no longer able to earn an income washing clothes. The loss of income means she can no longer provide her children with a balanced diet. Netsa’s mother worries about her children’s future. She says, ‘As I told her about my health status, she is thinking about it and her future too and I am also worried about her.’ Netsa’s mother now relies on her own mother for financial support and the family lives with her uncle. She is very concerned about what might happen to her children should she die. Children who have a family member experiencing long-term illness position themselves lower on the subjective well-being ladder, and those who have lost their father rate their own health as lower. Therefore, while our focus is the impact of risk and protective processes on children’s life chances, ill-health and illness need to be examined at the household level. This builds up a more complete picture, as not only does illness of one member impact on all the others, the entire household including children is involved in managing ill-health, including intergenerational dependence, as illustrated by Netsa’s family. 5.3 Children and households depend on multiple sources of informal and formal social support in responding to illness Children and families depend on multiple sources of support to manage ill-health and mitigate the impact of illness on the household. Children and caregivers employ a range of strategies. These may include reducing consumption, selling assets or using savings (consumption smoothing), labour substitution (by children or others), borrowing, seeking assistance from relatives, neighbours or NGOs, and/or use of social protection schemes. Families often employ several strategies simultaneously, as illustrated by Haymanot’s case. Households depend on multiple sources of support in coping with illness Haymanot lives in rural Tigray, Ethiopia. Her single mother has been very ill through heart problems. While Haymanot’s mother is still able to do some work from home, she cannot go to market and the household income has decreased. According to Haymanot: ‘We used to plough our plot by paying daily labourers but now we have no money and we use sharecropping.’ Consequently, her family currently receives only half of the output from the land. They have less food than before: ‘We used to eat bread and drink tea but now we only eat injera and sometimes kolo.’ Haymanot’s mother has visited a traditional healer for ‘cutting treatment’ on three occasions at a cost of 10 birr per visit. In order to ease the financial burden, Haymanot and her sister participate in the safety net programme by watering trees, digging and building fences; together they earn 20kg of grain each month.10 Haymanot has ceased schooling as a result of her mother’s illness, a situation which makes her unhappy: ‘I feel bad because I have to work and my mother is sick ... I will be happy if I continue going to school and my mother gets better.’ www.younglives.org.uk/what-we-do/our-policy-work Young Lives Policy Paper 4: September 2011 Page 23 Coping strategies are dependent on a range of factors including socio-economic status, household composition, social networks, the season, and geographic location (Okamoto 2011). In the qualitative interviews children and caregivers describe an order of coping strategies, starting with using savings (if this is an option) followed by seeking gifts or loans from relatives, then informal credit from money-lenders or low-interest loans from semi-informal groups (such as the self-help groups in Andhra Pradesh or burial societies called idirs in Ethiopia) and finally, again if possible, credit from formal institutions. In the household survey in Andhra Pradesh caregivers were asked to whom they would turn if they needed to raise 1,000 rupees in a week, and 62 per cent said they would turn to relatives and friends in the same town. As explained by Rajesh in Andhra Pradesh: ‘If we have money we don’t ask anyone. If not we ask our relatives, if they too don’t have, we ask other people in the village.’ However, this is problematic in communities where all households lack resources.10 Informal sources of support include both income-pooling within the household, or incomesharing between households in the extended family, and loans from individuals with a higher socio-economic status. Assistance can be financial, to pay for healthcare for example, or in the form of labour around the house or on the land. Caregivers stress the reciprocal nature of this type of support. Informal social support is based on reciprocity Ramya is from a rural Backward Caste community in Andhra Pradesh. Her father was suffering from kidney stones and had to go to Hyderabad for treatment as the local hospital did not have appropriate surgical facilities. Her family needed to pay 70,000 rupees for his treatment in advance. This was made possible by an interest-free loan of 100,000 rupees from Ramya’s maternal grandfather and uncle. According to Ramya’s mother, ‘of course it had to be returned. We help them when they need.’ Informal sources of social support also include loans or credit provided by individuals such as petty traders, teachers, and landowners. As a caregiver in Ethiopia explains: ‘There is no organisation that gives loans. There is a shop owner nearby. I took a loan from him with interest. I took 100 birr with interest.’ However, this also depends on social standing within the community, as one caregiver in Ethiopia explains: ‘Only those who are trusted by the kebele leaders can receive credit.’ While different households follow different pathways, typically, after exhausting informal sources of support many families turn to semi-informal sources, here meaning communitybased groups with clear structures and organisation. In Ethiopia these include iqub (savings and credit associations) and idirs (funeral associations in which members make regular payments and payouts are made to assist members with the cost of funeral, and at other times of need) (Dercon et al. 2006; Aredo 1993). In Andhra Pradesh self-help groups, known as Velugu or Indira Kranthi Pathakan (IKP), total over 700,000. These groups have been funded by national and state-level government, NGOs and international donors, including the World Bank. Promoted under the Development of Women and Children in Rural Areas (DWCRA) programme the aim has been to build on existing community groups by linking them with credit institutions to facilitate poverty reduction and economic development. The programme has focused on encouraging women to join in order to facilitate their economic empowerment (Galab and Chandrasekhara Rao 2003). 10 It is illegal for people under 18 years old to participate in the Ethiopian Productive Safety Net Programme (PSNP) or the National Rural Employment Guarantee Scheme (NREGS) in Andhra Pradesh. However, children describe using family members’ job cards to go and work on the schemes (Camfield and Vennam forthcoming). www.younglives.org.uk/what-we-do/our-policy-work Page 24 Poverty, Risk and Families’ Responses Regression analysis controlling for other factors shows household membership of community groups is associated with better objective and subjective outcomes for children. While this is not necessarily causative, groups may provide more social and economic support and so enable households to respond better to shocks. ●● In Andhra Pradesh low household group membership is associated with poorer child outcomes: children who live in households with low group membership are half as likely to be enrolled in school. In Ethiopia children who live in households which belong to self-help groups are 2.2 times more likely to be enrolled. In families which have no or little access to social support groups, children are 40 per cent less likely to have a healthy BMI-for-age in Andhra Pradesh, 40 per cent less likely to have a healthy height-for-age in Ethiopia and Peru, rate themselves lower in subjective wellbeing in Ethiopia and health in Andhra Pradesh. ●● ●● Semi-informal sources of social support can therefore assist families affected by illness, as illustrated by Denbel’s case. Informal sources of social support helps families cope better with shocks Denbel lives in urban Ethiopia with his single mother and younger brother. Denbel had malaria and he was taken to a private clinic, tested and provided with medication. The total cost of the treatment was 30 birr, which Denbel’s mother ‘borrowed ... with interest from our association in which we save money for buying meat for holidays’. The daily amount of interest on 30 birr is 90 cents and Denbel’s mother repaid half of the loan within a week, reducing the daily interest to 45 cents. A neighbour provided Denbel with support during his recovery: ‘The doctor told him to take the medication with milk. And [the] neighbour, who owns a cow, gave him one bottle of milk every day for him.’ As part of a qualitative sub-study on the National Rural Employment Guarantee Scheme (NREGS) programme in Andhra Pradesh, focus group discussions were held with children and caregivers. One group of mothers from a rural area in Anantapur district discussed the different formal and informal schemes in the community. On the one hand self-help groups were described as a ‘big headache’ as ‘when we do not have work then how can we repay the amount to the self-help groups’. As well as the social humiliation from peers at not being able to repay, defaulting on payments meant members would not qualify for the lower interest rate available to members. On the other hand, caregivers felt that self-help groups were a useful resource, enabling them ‘to meet the expenses in simple weekly instalments without the drudgery of heavy interest’. The ‘mutual understanding’ of self-help groups was preferred to taking ‘credit from the landlords in times of need’ as ‘we again fall into the circle of debts’. It was noted that it is difficult to obtain large sums of money, beyond 200–300 rupees, from self-help groups; caregivers explained the difficulties of accessing loans, as this is dependent ‘on the repayment capability of the person who has taken the loan’. Factors influencing the likelihood of obtaining a loan include the number of sons the household has; if a household has more sons then there is more inclination to lend money as the family send the sons to work and can repay the amount without fail. As is the case with households who own individual assets like land, there are more chances for recovery of the money. In contrast, households with school-age children are less likely to get loans. www.younglives.org.uk/what-we-do/our-policy-work Young Lives Policy Paper 4: September 2011 Page 25 While these examples illustrate the importance of semi-formal sources of support, they also highlight that these do not provide protection for everyone. Informal support is likely to be most available to those with some resources and greater social connectedness in the community. Moreover, families worry about their ability to pay back loans: ‘I don’t take Iqub. You know why? A person can take Iqub if he can pay back. You need to have a business or good income. It needs good income to participate in Iqub’ (caregiver, rural Ethiopia). This highlights the potential role for social protection, alongside existing structures, acting as an insurance or redistribution system, rather than a credit system, as illustrated by the following example from Andhra Pradesh. Formal and informal sources of social support can work together to mitigate the impact of shocks on families Rajesh is from a Scheduled Tribe and lives in rural Andhra Pradesh. His father contracted malaria and typhoid. ‘We thought he was going to die’, Rajesh recalls. The severity of the illness meant that Rajesh’s father was unable to work for a month. According to Rajesh, before his father fell ill ‘the money was okay [but] now it is horrible’. Faced with destitution, the family adopted a number of strategies. Rajesh and his siblings felt compelled to take work in the National Rural Employment Guarantee Scheme and in agriculture. Rajesh did not have his own work card, so used his sister’s during the holidays. Rajesh’s elder sister also sought to help the family by providing 500 rupees for medical costs and his mother borrowed 1,500 rupees, with one per cent monthly interest, from a self-help group. Social protection provides households with more options, but without the danger of debt traps. Data from Young Lives show that formal systems of social protection sit alongside, rather than displacing informal and semi-informal sources of support, as demonstrated by Rajesh’s family. 5.4 Access to quality healthcare can strengthen protection for children The MDGs include the goal of halving the mortality rate of children under 5 years of age. This focus is essential not only because this is the age when the risk of mortality is greatest, but also because a focus on survival impacts health and nutrition positively, with important gains in terms of children’s development more generally. However, as indicated, it is not just illness in the early years that has a profound impact on children. Orkin’s (2011) research in Ethiopia found that serious illness had the same negative impact as stunting on children’s schooling. Access to preventative and curative measures is therefore essential in reducing the burden of illness among children and family members. The qualitative interviews with children and caregivers reveal three different pathways in responding to illness. Typically, the first pathway involves treatment at home and then, if it persists, seeking treatment elsewhere: First my parents put ice and alcohol on the sore. I was treated in this way for one month. However, I was seriously sick and I was taken to the modern health centre. I had one medicine by injection and another medicine which was taken in the form of a fluid. Then I was able to recover from the injury (Habtamu, aged 13, rural Ethiopia). In the second pathway of seeking treatment there are three options; sometimes these are pursued concurrently. One option is to seek traditional healing practices or medicines. Some illnesses are believed to be treatable only by traditional medicine; as one caregiver in Ethiopia explained, ‘this disease does not like the modern medicine’. This is often the case for illnesses which are attributed to curses or spirit possession. Traditional medicine is not, however, necessarily a cheaper option. In Andhra Pradesh a caregiver explained how her daughter, who is the older sister of one of the children in the Older Cohort, suffers from problems ‘due to devils’. www.younglives.org.uk/what-we-do/our-policy-work Page 26 Poverty, Risk and Families’ Responses Her son-in-law spent between 7,000 and 8,000 rupees on traditional healing at the temple and for healers to visit the home. However, she says ‘she falls unconscious… lot of mystic talisman tied… we spent Rs.2000, Rs.1500 or Rs.1000 wherever we went… still they did not help her recover’. She was also taken to a doctor and had her blood and urine tested, received injections, drops and glucose tablets, all of which cost 700 rupees; but still she has not recovered. The second option consists of NGO-run services, which are either free or cost less than the third option, private clinics. In Ethiopia, 94.5 per cent of Young Lives households stated that they had to pay for medical treatment at a public facility in the case of a severe illness or injury (Barnett and Tefera 2010). The poorest households either did not receive exemptions or the exemptions did not cover the entire fees. Consequently, the third and final pathway is taking no direct action because of barriers, either financial, or distance from health facilities. One caregiver in rural Ethiopia explained how her son is suffering from trachoma and even though she took him to a clinic ‘the cost is very expensive and I couldn’t pay; so he didn’t get medication’. To address these issues the Government of Ethiopia introduced the ‘Accelerated Expansion of Primary Healthcare Coverage’ programme in 2003, focusing in particular on maternal and child health. The programme established health posts staffed by two trained healthcare extension workers in each rural kebele. As a result the percentage of the population covered by primary health services has increased substantially, from 51 per cent in 1998 to about 90 per cent in 2008 (Woldehanna et al. 2011). This expansion of services is reflected in Young Lives data where very few households were covered in 2006, compared with 27 per cent in 2009. Yet this expansion appears uneven since coverage is greater for households that are non-poor, male-headed and with higher levels of maternal education, than for those that are poor, female-headed and with lower levels of maternal education, mainly in rural areas. The results of the Health Extension Services have been impressive in coverage terms, but more remains to be done. Ethiopia is making good progress on MDG 4, as under-5 mortality has declined from 164 per 1,000 live births in 2005 to 104 per 1,000 live births in 2009 (Save the Children 2011). At the same time concerns have been raised about the low quality of the services, especially in reducing incidence of childhood diarrhoea (Admassie et al. 2010). While child mortality has declined, the proportion of children who are deprived of basic healthcare significantly increased between 2000 and 2005 (Save the Children 2011), and it has been estimated that 38 per cent do not receive routine injections or treatment for diarrhoea. Moreover, the same research showed that children in the poorest wealth quintile were twice as likely to be deprived of healthcare as children in the top quintile. This suggests that gains in healthcare expansion have been made in the higher wealth quintiles, rather than among the poorest. The research also suggests that a similar trend of the poorest households being left behind in access to healthcare is occurring in India and Peru (Save the Children 2011). This suggests the need for greater attention to the reach of healthcare services for the poorest and most marginalised households, and to quality of treatment for common childhood illness such as diarrhoea. In the Young Lives qualitative data, concerns over the quality of healthcare were raised by caregivers in both Ethiopia and Andhra Pradesh, particularly within the public sector. As one caregiver from rural Ethiopia explains: In this study area there is one health centre [private] and there is another health post here [public]. It [health post] is not effective. The service is poor in finding malaria disease since www.younglives.org.uk/what-we-do/our-policy-work Young Lives Policy Paper 4: September 2011 Page 27 the health practitioners are not capable. Secondly, since there is one health centre we have to wait in line before early in the morning. … Moreover, it is very difficult to go to the health centre before early in the morning so it is very terrifying to walk through dark. Concerns over the time it takes to access healthcare and the quality of care were echoed by caregivers in Andhra Pradesh: ‘With private doctors they will tell you what the position is. Going to government hospitals means going here to there as they transfer you. Who will be at home, who will be with them [the sick person] and who will look after the cattle?’ Consequently people are forced to opt for private healthcare. Research on healthcare in India published in The Lancet reflected these concerns, stating that ‘in 2004 the most commonly cited reason why people did not avail themselves of public services was not being satisfied with medical treatment by government doctors or facilities’ (Kumar et al. 2011: 671). As a result, 78 per cent of the population in rural areas and 81 per cent in urban areas access private outpatient care. However, concerns have been expressed over the lack of regulation and unreasonably high costs of these facilities (ibid.). The impact of private healthcare costs on families is demonstrated by the following case. Private healthcare costs creates large household debts Ravi is a member of the Scheduled Castes who lives in rural India. His mother suffers from stomach pains and headaches and also from the chikungunya virus which has so far resulted in treatment costs of around 6,000–7,000 rupees. Whenever Ravi’s mother has been unwell, his family has had to take out loans to pay her medical bills. She consults private doctors because she believes ‘government doctors will not care’; each consultation costs 400–500 rupees. The family currently owes 20,000 rupees to a local family and 8,000 rupees to the DWACRA micro-credit group. According to Ravi’s mother, ‘when we get nice crop, then we should repay ... [I]f we have to eat or not, we should repay the loans.’ Ravi has tried to supplement the family’s farming income by engaging in other labour activities: ‘I went [to another town] only to make money as we need it to repay our debts and also to run our household.’ Caregivers talk about having to delay seeking treatment while they find the money required for fees. The following is an example from Ethiopia: Caregiver: I have to borrow money to take them [children] for treatment. It takes two or three days before I get the money. Interviewer: So, can’t you take your children for treatment on time? Caregiver: No I can’t. Delays in seeking treatment because of financial or other barriers could have potentially serious consequences for children by reducing the success of treatment or increasing its cost. In Andhra Pradesh a healthcare card scheme (Aarogyasri) has been introduced to increase access. However, this is for secondary care only and does not cover common but serious illness, such as malaria. A health insurance programme is also being introduced throughout India but it has not yet been rolled out in Andhra Pradesh. In contrast, although in the early stages, Vietnam has been implementing a health insurance scheme. In time it will be possible for Young Lives to examine the impact of these schemes on outcomes for children. In Vietnam the health insurance law came into effect in July 2009. Children under the age of 6, ethnic minorities and poor people living in rural areas are entitled to free healthcare. The HealthCare Financing Reform in Ethiopia is establishing a list of free services to be delivered by health facilities and to include fee-waivers for the poorest. However, targeting can be problematic and can introduce tensions into communities, as well as being costly in terms of administration www.younglives.org.uk/what-we-do/our-policy-work Page 28 Poverty, Risk and Families’ Responses (Barnett and Tefera 2010). Attention to financing needs to be accompanied by investment into health facilities and medicines in order to engage uptake of services. The following case not only highlights how rising household consumption levels result in better outcomes for children, but also illustrates a way in which education and health services can be linked up in order that progress in one outcome is not undermined by obstacles in another. Linking education and health services Biritu lives in rural Ethiopia and is 13 years old. She is the sixth child of seven. She describes how ‘in the past years my parents did not take us to the clinic when we got sick, but currently since my parents earn money they can take us to clinics as soon as we get sick’. Biritu describes how her learning is impacted by eyesight problems, a recurring theme in the qualitative interviews with children and caregivers in Ethiopia. She explains ‘When I look at the whiteboard tears drop out of my eyes. I experience headache when noisy voices disturb me.’ For the headaches she is able to get tablets from the clinic at school free of charge. Key messages This section explored risk and protective processes through the lens of illness. Ill-health and illness not only have a greater impact on poor households in terms of direct and indirect costs but also poor families have greater difficulties in accessing good-quality services. The findings illustrate the key themes of this paper: ●● Poor households are affected by multiple and recurrent experiences of illness which are associated with other economic and environmental shocks, particularly due to food shortages. This results in poorer objective and subjective outcomes for children. Illness plays a part in the intergenerational transmission of poverty, as coping strategies often result in exposure to further risk. Illness is both a cause and consequence of poverty, due to the poor-quality living environment and the indirect and direct costs of illness. The entire household, including children, is involved in managing ill-health, which is a considerable source of anxiety for poor families. Children and households need multiple sources of support to help manage risk and are heavily dependent on informal social support from family and friends. There are also semi-informal community-based organisations, and household membership of these groups is associated with better outcomes for children. However, there are also dangers associated with debt-traps, and ability to access credit is often dependent on the household’s social standing within the community. Social protection schemes can therefore play an important role in providing poor families with more options. Access to quality healthcare can strengthen protection for children but is shaped not only by financial and time cost implications, but also sociocultural beliefs and concerns over quality, particularly within the public sector. Use of the private sector can result in higher debt burdens for households and there are questions over regulation and the exacerbation of inequalities in terms of how these services can be accessed. ●● ●● ●● www.younglives.org.uk/what-we-do/our-policy-work Young Lives Policy Paper 4: September 2011 Page 29 6. Conclusions and policy discussion: from ‘beating the odds’ to ‘changing the odds’ As 2015, the end of the period set by world leaders for achieving the MDGs, approaches, questions are being asked concerning how far progress has been made and whether this has been inclusive and equitable (Kabeer 2010). There has been positive progress on indicators such as primary school enrolment, and it is important to build on this foundation by addressing more complex issues such as education quality and the root causes of poverty. Evidence from Young Lives presented in this paper demonstrates how children’s life chances are structured by experiences of poverty and inequalities that are often multiple and intersecting, so consistently disadvantaging the same children. Inequalities and associated poverty appear the greatest risk factors, which through the uneven distribution of resources and skills can also undermine resilience or protective processes (UNICEF 2011: 2). This final section reflects on these key findings and the implications for policy, encouraging arguing for a move from focusing on ‘beating the odds’, namely the individual characteristics of a child, to ‘changing the odds’, namely the broader structural factors and social processes that shape the life chances of children and their families (Seccombe 2002). ●● Economic growth is necessary for poverty reduction but needs more harnessing for improving children’s life chances Despite falls in absolute poverty, disparities between children living in urban and rural areas and between different ethnic and caste groups are significant and become wider over the life course, although the picture on gender is less clear. This suggests that GDP growth, while important, will not alone lead to poverty reduction and better life chances for all children. One of the criticisms of the MDGs has been that they could encourage a focus on ‘low-hanging fruit’ and do not monitor whether the benefits of development reach the most marginalised. Disaggregated data are important to understand who is benefiting from poverty reduction and whether this is translating into better outcomes for children. Introducing disaggregation by the bottom quintile of household consumption into whatever framework follows the MDGs would encourage policies to focus on inequalities within countries (Dornan 2010a; Vandemoortele 2011). Addressing poverty and inequalities requires greater linking between social and economic policies Chronic poverty is compounded by multiple and recurrent adverse events, such as illness, which have lasting impacts on children’s life chances. These multiple risks and their intersecting nature particularly suggest the need for social policies which do not target one risk or sector alone, especially in light of emerging complex threats such as climate change (Wachs forthcoming). Poor outcomes in one area, such as health, can undermine progress in another, such as education. Quality, reach and appropriateness of services are all important Chronic poverty and structural inequalities mean that the same children are repeatedly disadvantaged, doing less well across a series of indicators in education, health and wellbeing. More attention is required on the reach and quality of services to ensure the most marginalised have access to good-quality services. This includes careful monitoring of ●● ●● www.younglives.org.uk/what-we-do/our-policy-work Page 30 Poverty, Risk and Families’ Responses the impact of the private sector, whether in education or in health, which appears to be exacerbating inequalities and placing additional burdens on the poorest households. It requires also that service design takes into account the needs and perspectives of users in order to build systems which are flexible and responsive. ●● Social protection schemes can provide poor families with more options and do not necessary replace systems of informal social support Social protection, informal social support and access to quality services can all be important sources of protection and support for children and households. Having more options enables families to mitigate the impact of shocks on the household. Perspectives from children and their caregivers are vital in understanding how these schemes and networks do and do not work and should inform the strengthening of existing systems and point to the gaps. Young Lives data show the linkages between child outcomes, household situations and broader structural factors and social processes. Concepts of risk and protection and their application to policy should shift from a focus on enabling individual children to ‘beat the odds’ to ‘changing the odds’, targeting the root causes of children’s poor life chances, namely poverty and inequalities, rather than just the symptoms. References Admassie, A., D. Abebaw and A. Woldemichael (2010) ‘Tackling Poor Healthcare: An Evaluation of Ethiopia’s Health Services Extension Programme’, New Delhi: Global Development Network Anthony, E.J. (1987) ‘Risk, Vulnerability, and Resilience: An Overview’ in E. J. Anthony and B.J. Cohler The Invulnerable Child, New York: Guildford Press Appadurai, A. (2004) ‘The Capacity to Aspire: Culture and the Terms of Recognition’ in V. Rao and M. Walton (eds.) Culture and Public Action, Palo Alto CA: Stanford University Press Aredo, D. (1993) The Informal and Semi-Formal Financial Sectors in Ethiopia: A Study of the Iqqub, Iddir, and Savings and Credit Cooperatives, Research Paper 21, Nairobi: Economic Research Consortium Barnett, I. and B. Tefera (2010) Poor Households’ Experiences and Perception of User Fees for Healthcare: A Mixed-Method Study from Ethiopia, Working Paper 59, Oxford: Young Lives Barton, W. (2005) ‘Methodological Challenges in the Study of Resilience’ in M. Ungar Handbook for Working with Children and Youth: Pathways to Resilience Across Cultures and Contexts, Thousand Oaks, Sage Publications Boyden, J. (2009) ‘Risk and Capability in the Context of Adversity: Children’s Contributions to Household Livelihoods in Ethiopia’, Children, Youth and Environments 19.2: 111–37 Boyden, J. and E. Cooper (2009) ‘Questioning the Power of Resilience: Are Children Up To the Task of Disrupting the Transmission of Poverty?’ in J. Addison, D. Hulme and R. Kanbur (eds.) Poverty Dynamics: Measurement and Understanding from an Interdisciplinary Perspective, Oxford: Oxford University Press Boyden, J. and G. Crivello (2011, forthcoming) ‘Growing up in Poverty: Politics and Perception as Mediators of Childhood Risk: The Case of Andhra Pradesh’ in J. Boyden and M. Bourdillion (eds.) Childhood Poverty: Multidisciplinary Approaches, Basingstoke: Palgrave Macmillan Boyden, J., A. Pankhurst and Y. Tafere (forthcoming) ‘Harmful Traditional Practices and Child Protection: Contested Understandings and Practices of Female Early Marriage and Female Genital Cutting in Ethiopia’, Development in Practice, London: Routledge Camfield, L. and U. Vennam (forthcoming) ‘Security, Productivity and “Social Equity”? A Community- www.younglives.org.uk/what-we-do/our-policy-work Young Lives Policy Paper 4: September 2011 Page 31 Level Exploration of Tensions Within the Mahatma Gandhi National Rural Employment Guarantee Scheme’, Young Lives Working Paper, Oxford: Young Lives Chuma, J., L. Gibson and C. Molyneux (2007) ‘Treatment-Seeking Behaviour, Cost Burdens and Coping Strategies Among Rural and Urban Households in Coastal Kenya: An Equity Analysis’, Tropical Medicine and International Health 12(5): 67388 Crivello, G. (2011) ‘Becoming Somebody’: Youth Transitions through Education and Migration in Peru’, Journal of Youth Studies 14.4: 395 411 Crivello, G. and J. Boyden (2011) Situating Risk in Young People’s Social and Moral Relationships: Young Lives Research in Peru, Working Paper 66, Oxford: Young Lives Crivello, G. and N. Chuta (2012, forthcoming) ‘“I Look for Help from Others When I Can’t do Things on my Own”: Sources of Support for Orphaned and Vulnerable Children in Ethiopia’, Development in Practice, London: Routledge Cueto, S., J. Escobal, M. Penny and P. Ames (2011) Tracking Disparities: Who Gets Left Behind? Initial Findings from Peru, Young Lives Round 3 Survey Report, Oxford: Young Lives Cueto, S., J. Leon, G. Guerrero and I. Muñoz (2009a) Psychometric Characteristics of Cognitive Development and Achievement Instruments in Round 2 of Young Lives, Technical Note 15, Oxford: Young Lives Cueto, S., G. Guerrero, J. León, E. Seguin and I. Muñoz (2009b) ‘Explaining and Overcoming Marginalisation in Education: A Focus on Ethnic/Language Minorities in Peru’, background paper for EFA Global Monitoring Report 2010, Paris: UNESCO Dercon, S. (2008) Children and the Food Price Crisis, Policy Brief 5, Oxford: Young Lives Dercon, S., J. De Weerdt, T. Bold and A. Pankhurst (2006) ‘Group-based Funeral Insurance in Ethiopia and Tanzania’, World Development 34(4): 685703 Dercon, S. and A. Singh (2011, forthcoming) ‘From Nutrition to Aspirations and Self-Efficacy: Gender Bias over Time among Children in Four Countries’, background paper prepared for World Bank World Development Report 2012: Gender Equality and Development, Washington DC: World Bank Dornan, P. (2010a) Children and the Millennium Development Goals: Fragile Gains and Deep Inequalities, Policy Paper 2, Oxford: Young Lives Dornan, P. (2010b) Understanding the Impact of Crisis on Children in Developing Countries: Round 3 Preliminarily Findings, Oxford: Young Lives Dornan, P. (2011) Growth, Wealth and Inequality: Evidence from Young Lives, Policy Paper 5, Oxford: Young Lives Feinstein, L. (2003) ‘Inequality in the Early Cognitive Development of British Children in the 1970 Cohort’, Economica 70: 7397 Galab, S. and N. Chandrasekhara Rao (2003) ‘Women’s Self-Help Groups, Poverty Alleviation and Empowerment’, Economic and Political Weekly 127483 Galab, S., S.V. Kumar, P.P. Reddy, R. Singh and U. Vennam (2011) The Impact of Growth on Childhood Poverty in Andhra Pradesh: Initial Findings from India, Young Lives Round 3 Survey Report, Oxford: Young Lives Harper, C., R. Marcus and K. Moore (2003) ‘Enduring Poverty and the Conditions of Childhood: Lifecourse and Intergenerational Poverty Transmissions’, World Development 31(3): 535–54 Hart, J. (2008) Business as Usual? The Global Political Economy of Childhood Poverty, Technical Note 13, Oxford: Young Lives Himaz, R. (2009) The Impact of Parental Death on Schooling and Subjective Well-being, Working Paper 44, Oxford: Young Lives Jones, N., C. Harper and C. Watson (2010) Stemming Girls’ Chronic Poverty: Catalysing Development Change by Building Just Social Institutions, Manchester: Chronic Poverty Research Centre Kabeer, N. (2000) ‘Inter-generational Contracts, Demographic Transitions and the ‘Quantity-Quality’ Tradeoff: Parents, Children and Investing in the Future’, Journal of International Development 12: 46382. www.younglives.org.uk/what-we-do/our-policy-work Page 32 Poverty, Risk and Families’ Responses Kabeer, N. (2010) Can the MDGs Provide a Pathway to Social Justice? The Challenges of Intersecting Inequalities, New York: UNDP Krutikova, S. (2009) Determinants of Child Labour: The Case of Andhra Pradesh, Working Paper 48, Oxford: Young Lives Kumar, A.K.S., L.C. Chen, M. Choudhury, S. Ganju, V. Mahajan, A. Sinha and A. Sen (2011) ‘India: Towards Universal Health Coverage 6: Financing Health Care for All: Challenges and Opportunities’, Lancet 377: 66879 Le Thuc, D. (2009) The Effect of Early-age Stunting on Cognitive Achievement among Children in Vietnam, Working Paper 45, Oxford: Young Lives Le Thuc, Duc, Nguyen Thang, Nguyen Van Tien, Mai Thuy Hang, Vu Thi Thu Thuy (2011) How Do Children Fare in the New Millennium: Initial Findings from Vietnam, Young Lives Round 3 Survey Report, Oxford: Young Lives Levine, R., C. Lloyd, M. Greene and C. Grown (2008) Girls Count: A Global Investment and Action Agenda, Washington: Center for Global Development Lloyd, C. with J. Young (2009) New Lessons: The Power of Educating Adolescent Girls, Washington: Center for Global Development Luthar. S., D. Cicchetti and B. Becker (2000) ‘The Construct of Resilience: A Critical Evaluation and Guidelines for Future Work’, Child Development 71, 543–62 Maplecroft (2011) Climate Change Vulnerability Index ‘Big economies of the future - Bangladesh, India, Philippines, Vietnam and Pakistan - most at risk from climate change’, http://maplecroft.com/about/news/ ccvi.html Masten, A., K. Best, N. Garmezy (1990) ‘Resilience and Development: Contributions from the Study of Children Who Overcome Adversity’, Development and Psychopathology 2: 425–44 McIntyre, D., M. Theide, G. Dahlgren and M. Whitehead (2006) ‘What are the Economic Consequences for Households of Illness and of Paying for Health Care in Low- and Middle-Income Country Contexts?’, Social Science and Medicine 62: 85865 Morrow, G. and U. Vennam (2012, forthcoming) ‘How Children Manage Risk in Daily Life: Identifying Hazards in Agricultural Work in Andhra Pradesh, India’, Development in Practice, London: Routledge Okamoto, I. (2011) ‘How do Poor Rural Households in Myanmar Cope with Shocks? Coping Strategies in a Fishing and Farming Village in Rakhine State’, The Developing Economies 49.1: 89112. Orkin, K. (2011) ‘If God wills…next year I will send her back to school’: The Effects of Child and Parental Illness on School Participation in Rural Ethiopia, CREATE Pathways to Access Research Monograph 60, Falmer: University of Sussex Outes, I., A. Sanchez and O. Molinas (2010) Psychosocial Status and Cognitive Achievement in Peru, Working Paper 65, Oxford: Young Lives Pells, K. (2010) Inequalities, Life Chances and Gender: Round 3 Preliminarily Findings, Oxford: Young Lives Pells, K. (2011) Poverty and Gender Inequalities: Evidence from Young Lives, Policy Paper 3, Oxford: Young Lives Rojas, V. (2011) ‘Grades are Sacred… We’d Rather be Beaten with a Stick’: Students’ Perceptions of Discipline and Authority in a Public High School in Peru, Working Paper 70, Oxford: Young Lives Rutter, M. (1985) ‘Resilience in the Face of Adversity. Protective Factors and Resistance to Psychiatric Disorder’, British Journal of Psychiatry 147: 598611 Sakthivel, S. and P. Joddar (2006) ‘Unorganised Sector Workforce in India: Trends, Patterns and Social Security Coverage’, Economic and Political Weekly May 27: 210714 Sanchez, A. (2009) Early Nutrition and Cognitive Achievement in Pre-School Children in Peru, Working Paper 57, Oxford: Young Lives Sauerborn, R., A. Adams and M. Hien (1996a) ‘Household Strategies to Cope with the Economic Costs of Illness’, Social Science and Medicine 43.3: 291301 www.younglives.org.uk/what-we-do/our-policy-work Young Lives Policy Paper 4: September 2011 Page 33 Sauerborn, R., A. Nougtara, M. Hien and H.J. Diesfeld (1996b) ‘Seasonal Variations of Household Costs of Illness in Burkina Faso’, Social Science and Medicine 43.3: 28190 Save the Children (2011) Healthcare Deserts: Severe Healthcare Deprivation Among Children in Developing Countries, London: Save the Children Schoon, I. (2006) Risk and Resilience: Adaptations in Changing Times, Cambridge: Cambridge University Press Seccombe, K. (2002) ‘“Beating the Odds” Versus “Changing the Odds”: Poverty, Resilience, and Family Policy’, Journal of Marriage and Family 64: 38494 Smart, R. (2003) Policies for Orphans and Vulnerable Children: A Framework for Moving Ahead, Washington DC: Policy Project Streuli, N., U. Vennam and M. Woodhead (2011) Increasing Choice or Inequality? Pathways through Early Education in Andhra Pradesh, India, Studies in Early Childhood Working Paper 58, The Hague: Bernard van Leer Foundation Temin. M. and R. Levine (2009) Start with a Girl: A New Agenda for Global Health, Washington: Center for Global Development UNESCO (2010) Reaching the Marginalised: EFA Global Monitoring Report 2010, Paris: UNESCO UNICEF (2010) State of the World’s Children 2011: Adolescence – An Age of Opportunity, New York: UNICEF UNICEF (2011) 2011 UNICEF Humanitarian Action for Children Building Resilience, New York: UNICEF Vandemoortele, J. (2011) ‘The MDG Story: Intention Denied’, Development and Change 42.1: 121. Vennam, U., G. Crivello and H. Kongara (2010) Children’s Perspectives on Risk and Vulnerability in Contexts of Poverty and Change: Findings from Andhra Pradesh, Working Paper 64, Oxford: Young Lives Wachs, T. (forthcoming) ‘Poverty, child risk and resilience in developing countries’ in J. Boyden and M. Bourdillon (eds.)s Childhood Poverty Multidisciplinary Approaches, Basingstoke: Palgrave Macmillan Werner, E. and R. Smith (2001) Journeys from Childhood to Midlife: Risk, Resilience and Recovery, Ithaca: Cornell University Press Woldehanna, T., R. Gudisa, Y. Tafere and A. Pankhurst (2011) Understanding Changes in the Lives of Poor Children: Initial Findings from Ethiopia, Young Lives Round 3 Survey Report, Oxford: Young Lives World Bank (2006) World Development Report 2007: Development and the Next Generation, Washington, DC: World Bank World Bank (2011) World Development Report 2012: Gender Equality and Development, outline report, available at <https://www-wds.worldbank.org/external/default/WDSContent Server/IW3P/IB/2010/11/03/0 00334955_20101103062028/Rendered/PDF/576270WDR0SecM1e0only1910BOX353773B.pdf> World Health Organization (2009) Protecting health From Climate Change: Global Research Priorities, Geneva: WHO World Health Organization (2010) World Health Report Year: Health Systems Financing. The Path to Universal Coverage, Geneva: WHO www.younglives.org.uk/what-we-do/our-policy-work © Young Lives 2011