Report No 48627-TR Turkey: Expanding Opportunities for the Next Generation A Report on Life Chances February, 2010 World Bank Europe and Central Asia Region Human Development Department Report No 48627-TR Turkey: Expanding Opportunities for the Next Generation A Report on Life Chances February, 2010 World Bank Europe And Central Asia Region Human Development Department World Bank Copyright @ 2009 The International Bank for Reconstruction and Development The World Bank 1818 H Street, NW Washington, DC 20433, USA All rights reserved The World Bank enjoys copyright under protocol 2 of the Universal Copyright Convention. This material may nonetheless be copied for research, educational or scholarly purposes only in the member countries of The World Bank. Material in this report is subject to revision. Currency Equivalents (Exchange Rate Effective December 11, 2009) Currency Unit = TL (Turkish Lira) US$ 1 = 1.49 TL ABBREVIATIONS AND ACRONYMS AÇEV Anne Çocuk Eðitim Vakfý (Mother Child Education Foundation) BADEP Baba Destek Programý (Father Support Program) CCT Conditional Cash Transfer ECD Early Childhood Development ERI Education Reform Initiative GDP Gross Domestic Product HBS Household Budget Survey HINOG High Intergenerational Opportunity Group KEDV Kadýn Emeðini Deðerlendirme Vakfý (Foundation for Support of Women’s Work) LINOG Low Intergenerational Opportunity Group METU Middle East Technical University MOCEP Mother Child Education Program MONE Ministry of National Education NGO Non-governmental organization OECD Organization for Economic Co-Operation and Development PISA Program for International Student Assessment PRONOEI El Programa No Escolarizado de Educacion Inicial (Non-School Initial Education Project) SHÇEK Sosyal Hizmetler Çocuk Esirgeme Kurumu (Social Services and Child Protection Agency) SPO State Planning Organization SSI Social Security Institution SYDGM Sosyal Yardýmlaþma ve Dayanýþma Genel Müdürlüðü (General Directorate of Social Assistance and Solidarity) TAPF Türkiye Aile Saðlýðý ve Planlamasý Vakfý (Turkish Family Health and Planning Foundation) TDHS Turkey Demographic and Health Survey TEÇGE Türkiye'de Erken Çocukluk Geliþim Ekolojileri (Study of Early Childhood Developmental Ecologies in Turkey) TL Turkish Lira TÜ�K Türkiye �statistik Kurumu (Turkish Statistical Institute) UHI Universal Health Insurance UNICEF United Nations Children’s Fund UNDP United Nations Development Program WDI World Development Indicators Vice President : Philippe H. Le Houerou Country Director : Ulrich Zachau Sector Director : Tamar Manuelyan Atinc Task Team Leader : Jesko Hentschel Turkey: Expanding Opportunities for the Next Generation Contents Forward and Acknowledgements ...................................................................................................................... v Overview ......................................................................................................................................................... vii 1. Introduction: Life Chances ........................................................................................................................... 1 2. Equity Aspirations ......................................................................................................................................... 2 3. The State of Equality of Opportunities in Turkey .........................................................................................4 4. Child Development and Child Risks ........................................................................................................... 11 5. Expanding Opportunities for the Next Generation: Early Childhood Development Policies and Programs .............................................................................. 16 6. Reflections.................................................................................................................................................. 25 References ...................................................................................................................................................... 27 Technical Endnotes.......................................................................................................................................... 30 Graphs Graph A : Relative Risk of Child Poverty and Stunting, by Intergenerational Opportunity Group, 2004 Graph B : Per Capita Social Expenditures by Age Group in Turkey (Central Government, Excluding Social Security Contributions, 2008), ’000 TL Graph C : Aggregate Income and Poverty Reduction Effects of Increased Female Labor Force Participation Graph 1 : Under Five Mortality Rate Over Time (1960-2006) Graph 2 : Should the Gap Between the Rich and the Poor in Turkey be Reduced? Graph 3 : What is the Main Reason why there are some people in need in Turkey today? Graph 4 : Labor Force Participation Rates, by Gender Graph 5 : Wealth Distribution and Circumstances, 2004 Graph 6 : Inequalities of Wealth Opportunity in Turkey Graph 7 : Opportunity Profile for Least and Most Advantaged Population Groups, 2004 (percent) Graph 8 : Height for Age Measures for LINOG and HINOG Graph 9 : Relative Risk of Child Poverty and Stunting, by Intergenerational Opportunity Group, 2004 Graph 10 : Relative Risk of Poverty, by age group 2003 and 2006 Graph 11 : Distribution of Estimated Numbers of Additional Poor, by Age Group, 2009-2010 Graph 12 : Child Development Measurement Framework Graph 13 : Cognitive Development Inputs at Household Level for 36-47 month old children in Turkey Graph 14 : Cognitive Development Scores of 36-47 month old children in Turkey Graph 15 : Enrollment Age Profiles by Opportunity Group and Gender, (2004) Graph 16a : Under-5 Mortality Rate (latest available year, per 1,000 live births) Graph 16b : Births Attended by Skilled Staff (latest available year, percent) Graph 17 : Coverage of Various ECD programs in Turkey by Age Category Graph 18 : Pre-primary Gross Enrollment for 36-72 months, percent 2008 Graph 19 : Preschools and Daycare Access, by socioeconomic status and rural/urban location (2008) Graph 20 : Pre-School Enrolment and Human Development Index, by Province Graph 21 : Distribution of Social Expenditures in Turkey by Age Group (2008) Graph 22 : Per Capita Social Expenditures by Age Group in Turkey (Central Government, Excluding Social Security Contributions, 2008), ‘000 TL Graph 23 : Aggregate Income and Poverty Reduction Effects of Increased Female Labor Force Participation Tables Table 1 : Intergenerational Opportunity Groups of Children Table 2 : Intergenerational Opportunity Groups and Current Residence of Children, percent Table 3 : Intergenerational Opportunity Groups and educational attainment of grandparents and parents, percent Table 4 : Child Development Access and Outcome Indicators by Opportunity Status, 2004, percent Turkey: Expanding Opportunities for the Next Generation v Forward and Acknowledgements This Report is part of a larger Welfare and Social Policy Organization and its Undersecretary, Kemal Madenoðlu work agenda which the Turkish State Planning Organization as well as to the Director General of Social Policies, and the World Bank are carrying out collaboratively. The Yýlmaz Tuna, the division chief for Labor and Social work agenda includes the preparation of a number of Policies, �lyas Çelikoðlu and their teams. In particular, jointly conceptualized, and in part co-authored, analytical we benefited from the technical insights and guidance studies on topics ranging from examining the determinants of Hande Palancýoðlu, Alper Yatmaz, Yýldýz Yapar, of female labor force participation to analyzing poverty Serdar Polat and Gökhan Güder. Further, we obtained trends, and from examining equity determinants to inves- much technical support and advice from colleagues in tigating the link between poverty, employment creation the Department of Social Statistics, headed by Enver and growth. Further, the work agenda also comprises a Taþtý, at the Turkish Statistical Institute, as well as number of human development dialogues for which we colleagues in the Directorate General of Pre-School are inviting international experts and practitioners to share Education at the Ministry of National Education, the their experiences about social policy reforms with the Directorate General of Primary Health and the Mental Turkish Government and the wider academic and non- Health Unit at the Ministry of Health. We are also governmental public (www.iged-tr.net). grateful to Prof. Nazlý Baydar from Koç University and her research team for both feedback and access to their In the spirit of this dialogue and our collaborative research study on Early Childhood Development engagement with partners in Turkey, the main purpose Ecologies (TEÇGE). Reza Hossaini (UNICEF) provided of this Report is to contribute to the public debate on us with much support throughout the implementation equity and social policy in society. The Report aims to of the overall work agenda. We would also like to thank present a number of broad findings and reflections for colleagues in the Early Childhood Development that debate which could then be drawn on when specific community in Turkey, in particular Ayla Göksel at the policy reforms are discussed. Mother Child Education Foundation (AÇEV) and Þengül Akçar at KEDV for sharing field experiences and The Report was produced by a team coordinated by Jesko Hentschel and comprised of Meltem Aran, Raif providing valuable insights. Can, Francisco Ferreira, Jeremie Gignoux, Elif Yonca Yükseker, and Arzu Uraz (all World Bank). It also Special gratitude is due to Prof. Sencer Ayata from draws on a technical note by Martha Llanos (consultant) Middle East Technical University (METU) who, on Early Childhood Development policy in Turkey. throughout the entire process, provided us with continuous feedback and suggestions. We are grateful From within the World Bank, we benefited from critical to him for organizing a series of dialogue meetings reviews by Juan Diego Alonso, Diego Angel-Urdinola, within the Social Policy Graduate Program of the Christian Bodewig, Wendy Cunningham (peer reviewer), Graduate School of Social Sciences at METU and Ariel Fiszbein (peer reviewer), and Jaime Saavedra facilitating continuous interaction between this work (peer reviewer). Gordon Betcherman, Tamar Manuelyan program and the academic community. Finally, we Atinc and Ulrich Zachau provided overall support and much benefited from seminars held during the guidance during the preparation of the Report. preparation of the collaborative work program at Boðaziçi University, Koç University, METU, Sabancý Our thanks for the collaboration in the drafting of this, University as well as with colleagues from the United Report go foremost to our partners in the State Planning Nations system in Turkey. Turkey: Expanding Opportunities for the Next Generation vii Overview 1. From the Marmara earthquake and the 2001 widely today – and that the country could vastly improve financial crisis to the onset of the global economic its human and economic development potential by slowdown in mid-2008, Turkey recorded major social expanding the opportunities available to its next and human development progress. Under-five generation, particularly the most disadvantaged children. mortality rates continued their remarkable, decade-long decline, reaching 23.9 (per thousand live births) in 4. Transforming opportunities into achievements 1 2008. Net enrollment rates in secondary school, often depends on one’s own drive, effort and, at times, characterized as the Achilles heel of human development luck. But the opportunities themselves might be in the country, climbed steeply from 51 percent (2002) determined by factors that any single person can do to 59 percent (2008). Similarly, poverty had decreased absolutely nothing about: in which family one is born from 27 percent in 2002 to 19 percent in 2007 – and and where; the educational background of one’s parents further decreases have likely materialized until mid- and their wealth; the language spoken in the family; or 2008. Such poverty reduction was achieved not only one’s own gender. Such factors, or circumstances, as through the strong growth performance of the economy John Roemer calls them, are all independent of one’s but also due to a marked reduction in inequality in own choices and effort. society: between 2003 and 2006, consumption inequality 5. A society offering equal opportunities to its citizens declined by more than ten percent. would then be one in which all those circumstances, such as the socio-economic family background or birth 2. Such improvements, welcome as they certainly place, matter little or not at all in determining life are, still lag the aspirations of the Turkish people, chances. One, in which a girl born in a remote, poor as eloquently documented in a 2006 opinion survey. area of the country to parents without much formal This survey registered a strong preference for a more education would have the same chances to succeed in equitable society among the Turkish public. A full 85 life as a boy born in the center of Izmir to wealthy and percent of the adult population voiced that the gap educated parents. between the rich and the poor in society should be reduced – the single highest proportion in an international 6. The equity concept used here places much comparative assessment which included Eastern emphasis on opportunities rather than on outcomes 2 European and Central Asian countries. In the same per se. Equality of outcomes would imply that reading inquiry, two out of three Turkish adults responded that scores of all children at age fifteen should be equally the main reason for why poverty existed in the country high, that wealth be distributed homogenously, or that was ‘injustice in society’ and close to three quarters of life expectancy should be the same for everyone in all respondents voiced that the poor should not be held society. But such outcomes also depend to a significant accountable for their fate. extent on one’s own efforts, disposition or luck. Also, rewarding effort (to learn, study and work) is essential The State of Equality of Opportunities in Turkey in societies for innovation and advancement. 3. This Report examines the reality behind those 7. We find that circumstances are important in statements. It focuses on equality of opportunity in determining life chances in Turkey today. We present Turkish society today–most importantly on the life quantitative estimates of inequality of opportunities for chances of today’s children, the country’s future. Our two outcome dimensions: economic and educational results show that life chances of people in Turkey differ achievement. Economic opportunity is measured as 1 Preliminary results based on Turkey Demographic and Health Survey (2008). 2 Data from Life in Transition Survey (2006). See Ferreira et al (2009) and Zaidi et al (2009). Turkey: Expanding Opportunities for the Next Generation viii household wealth, while the opportunity for education measured by the education they attained roughly forty is assessed through qualitative achievement (performance or fifty years ago – retains a powerful link to the well- 4 in standardized tests). being and chances of their grandchildren today. 8. In examining the distribution of household 11. For illustration purposes, the report distin- wealth, in 2004, we find that at least one-third of the guishes two groups in the inter-generational oppor- wealth disparity is due to inequality of opportunity 3 tunity profile, according to spatial and educational in the country. The most important circumstances that characteristics: a Low Inter-generational Opportunity determine opportunities in Turkey are area of birth and Group (or LINOG) and a High Inter-generational parental education. These alone account for two-thirds Opportunity Group (or HINOG). of inequality of opportunity related to the wealth distribution in the country. Limited analyses for other countries exist that would allow us to compare Turkey Graph A: Relative Risk of Child Poverty and Stunting, by Inter-generational Opportunity Group, 2004 internationally – as a rough indication, the country appears to rank towards the more moderate end of Latin American countries, a continent that has long been highlighted for the persistence of inequalities. 9. A similarly strong link between circumstances and outcomes emerges when we look at educational achievement for Turkey’s 15 year olds, as measured by results of the 2006 PISA scores. Similar to the results regarding the distribution of wealth, we find that between a quarter and a third of overall educational inequality can be traced to underlying inequality of opportunities. As was the case for the wealth distribution, the socio-economic background of the families in which the teenagers grew up in accounts for the lion’s share Note: Risk relative to average incidence of all children age 5 and below Source: Staff calculations based on Turkey Demographic and Health Survey (2004) of such unequal opportunities. Spatial variables retain their importance, but mostly along the rural/urban divide, signaling inequalities in access to quality education in 12. Such inter-generational opportunity groups are the country. Gender, which is a key determinant of closely aligned with the well-being of present day enrollment, is not a significant correlate of achievement children. The most striking result is the close – that is: once they succeed in staying in school, girls relationship between child poverty, in 2004, and the do the same as or better than boys in the achievement inter-generational opportunity groups: child poverty in tests. the LINOG was, with 78 percent, almost twenty times higher than the HINOG group in which child poverty From Grandparents to Grandchildren barely reached 5 percent as depicted in Graph A. 10. Comparing Turkish society today with the social 13. Similarly, more than a quarter of children in the and economic life of four or five decades ago reveals LINOG showed signs of stunting–an indicator that that tremendous transformations have taken place measures the long-term effects of malnutrition – while –regarding urban and rural life and livelihoods alike. in the HINOG only four out of one hundred children But such transformations notwithstanding, this report showed such growth retardation. While iodine deficiency finds that the socio-economic status of grandparents– – a leading cause for mental retardation during child 3 Estimations based on the Turkey Demographic and Health Survey fielded between December 2003 and May 2004. While our analysis captures more than 80 percent of all Turkish households, the circumstance variables we are able to examine pertain to the ever-married women in the household only. See Technical Endnote 1. 4 Given the data source (see Technical Endnote 1), we examine in this report the relationship between maternal grandparent education and the outcomes of their grandchildren. Turkey: Expanding Opportunities for the Next Generation ix development – was present in more than 80 percent for had the highest poverty rates among all age groups with LINOG children, it was detected in about 20 percent more than one in four children being poor. More than for those in the HINOG group. forty percent, or over five million, of all poor people in Turkey in 2006 were children below fourteen years 14. Stunting and iodine deficiency are, according to of age. About 1.8 million of them were infants and cross-country research, among several leading acute young children under five. And while overall poverty risk factors that signal children not being able to reach rates had decreased between 2003 and 2006, children their full cognitive development potential. Hence, we benefited least from such improvements. According to find that lack of opportunity travels through generations, estimations conducted for this Report, children are also and now significantly influences the development the population group that is likely to make up the largest potential of disadvantaged children in Turkey today. share of those falling into poverty due to the economic slowdown Turkey is witnessing today. 15. Girls appear to be particularly impacted by such inter-generational transmission of opportunities. 19. Findings from early childhood development Within the Low Inter-generational Opportunity Group, research stress that it is the multi-dimensional lack young girls are significantly more likely to show early of opportunities that put children at highest risk of signs of malnutrition than boys: their stunting rate is, not reaching their development potential. We therefore with more than 30 percent, about a third higher than examine core child development access levels (input their male siblings (23 percent). Child development indicators) and outcomes, along both child poverty as trajectories continue to differ by gender: in the LINOG– well as inter-generational opportunity group dimensions. the low inter-generational opportunity group–the likelihood of girls in the ages of 7 to 15 being enrolled 20. With the strong inter-generational transmission was 68 percent, compared to around almost 90 percent of inequalities in society, it comes as little surprise for boys. Hence, breaking the inter-generational that we find core inputs into the child development transmission of inequity would have to place particular process to diverge quickly according to child poverty emphasis on supporting disadvantaged girls in Turkish and opportunity status. Across the board, poor children society today. in low opportunity settings show remarkably lower access indicators to basic health functions than other Child Development and Child Risks groups – and this starts already before the children are 16. Our finding that the inter-generational trans- born, with two-thirds of poor mothers in low opportunity mission of inequity is powerfully affecting Turkey’s settings not receiving a minimum of antenatal care youngest generation today points to the need for during their pregnancies. Similarly, more than 90 understanding of how children’s opportunities percent of poor children in the low opportunity setting develop from a young age, and whether there are do not command food with sufficient iodine supply and policy interventions that can contribute to reduce the four fifths do not receive the full set of six recommended impact of exogenous circumstances on life chances. immunizations before they complete one year of age. Similarly, new data from Koç University show that inputs for cognitive development of children differ 17. Given the very close mapping of the interge- strongly by socio-economic strata. nerational opportunity profile to child poverty, one important indicator of life chances for today’s 21. Differences in development outcomes according children is their poverty status. Poverty indeed is a to child poverty and opportunity groups emerge circumstance for children given that it is defined at the quickly as well. Low birth weight affects a quarter of household level and children do not contribute to the poor children in low opportunity settings, indicating income or asset envelope of households. constraints in fetal nutrition during a crucial period for brain development – and it is largely due to poor maternal 18. In 2006, child poverty was higher than for any nutrition and infections. Similarly, high stunting rates other age group in Turkey. Both younger (zero to are concentrated in poor children of the LINOG group, five years) and older (six to fourteen years) children albeit they are also significant for poor children in better Turkey: Expanding Opportunities for the Next Generation x opportunity settings–indicating that acute risk factors 25. One pro-equity policy which could break the can also develop in poor households with a better cycle of the inter-generational transmission of educational background of the parents. The ongoing inequity focuses on early childhood development research study from Koç University equally shows that (ECD). The importance of ECD, starting at the time cognitive development scores, an early outcome measure, when the child is not even born, is due to brain already diverge quickly and early according to socio- development in the first months and years, affecting economic status of the families in which the children physical health, learning, and social behavior throughout grow up. life. 22. Later, these vastly different inputs and early As such, fostering ECD does not only attack poverty signals for divergence in outcomes feed through to but is also the key for tackling the inter-generational school attendance and completion –and thereby transmission of inequities. ECD programs aim to complete the cycle of the inter-generational transmission improve the survival, growth, and development of young of inequality of opportunity. Sharp contrasts emerge, children so that these can develop all the necessary both by opportunity group as well as, almost more cognitive, physical, and socio-emotional skills they pronounced, by gender. Such gender differences in need later on in life. education access, above and beyond the opportunity Graph B: Per Capita Social Expenditures by Age Group in Turkey (Central setting of households, once again emerges as one of the Government, Excluding Social Security Contributions, 2008), ’000 TL core development challenges for Turkey. 23. The above described core child development indicators as well as risk factors emphasize the importance the country now places on child development. This is also underlined if we compare Turkey internationally: bench- marking of indicators, such as immunization coverage, birth attended by skilled staff as well as outcome indicators such as the under-5 mortality rate, all show that Turkey, given its level of income, is somewhat underperforming in relation to comparator countries. Expanding Opportunities for the Next Generation Note: Staff calculations. (See Endnote 9 for explanations.) 24. Public policies that would reduce inequalities of opportunity in society are broadly those that would 26. Over the past years, Turkey has set itself ambitious attempt to weaken the link between circumstances, targets to raise core child development indicators. In which people cannot be held accountable for, and line with such targets, innovative and path-breaking outcomes. A wide array of such policies exist, starting reforms are being carried out that focus on children, including the pivotal role that family doctors are now from building human capital through providing skill playing in growth monitoring and assessing family support building and life-long learning opportunities for systems. disadvantaged groups to supporting productive asset creation by, e.g., addressing capital market imperfections 27. Few public resources reach children in Turkey or income transfers which would offset original today. While funding for Early Childhood Development disadvantage. Or these could be policies that open policies and programs has gradually increased in Turkey, opportunities by connecting people to markets, using we find that, on a per capita basis, only a small share public infrastructure investment to overcome of public funds, about 6.5 percent, is directed to children geographical poverty traps, and facilitating access for from zero to six years. Estimating central public budget people to move and benefit from higher opportunity expenditures (excluding contributions to the social areas. security system), on a per capita and age group basis, Turkey: Expanding Opportunities for the Next Generation xi we estimate that the population above 44, largely due to and changes in the returns to education or the demand high pension expenditures, receives a per capita transfer for labor. Hence, the effects could well be lower-bound at least two-and-a-half times as large as children today estimates with benefits even larger than reported here. (Graph B). 31. Second, some of these ‘investments in Turkey’s 28. Coverage, especially for disadvantaged children, future’ could very well materialize in the very short is generally low in most dimensions of ECD. Pregnancy run. A companion report by the World Bank and the monitoring and immunization show high coverage and State Planning Organization (World Bank 2009a) on about 30 percent of children 48-72 months of age are female labor force participation establishes that poor, enrolled in preschool. However, all other ECD programs especially urban women might not work, although they reach less than 10 percent. In addition, evidence from want to, because the very cost of childcare inhibits them multiple sources show that one of the most important from taking up income-earning activities. As a spin- programs – public preschool and day care – reaches many off to increased availability of pre-school and child care more children of wealthier than poorer families. Turkey services, a concomitant increase in female labor force has much to build on given that several of its civil society participation could hence achieve immediate growth and public initiatives have a proven track-record and impact and productivity effects: we estimate that an increase in reaching the poorest and most disadvantaged children in the female labor force participation rate to 29 percent early on. (to match the Government’s targets by 2013) could lead to a decrease in poverty by more than 15 percent and a possible aggregate income effects for all Turkish 29. Improving the opportunities of Turkey’s children households to be in the order of 7 percent – a substantial today would have significant economic and social increase that could significantly support households to benefits. Significant work has been carried out by “work themselves out of poverty” (Graph C). partner organizations to assess the potential costs of comprehensive ECD policies. Hence, this report focuses on the benefit side. Graph C: Aggregate Income and Poverty Reduction Effects of Increased Female Labor Force Participation 30. We use two simulation models to assess the 5 potential benefits of selected ECD policies. First, the Report uses the results of controlled experiments in Turkey that quantify the impact of pre-school/parenting on educational attainment of beneficiaries. We employ a model that examines how poverty, incomes, and female labor force participation would be different today for the generation of 20-39 year olds if they had attended pre-school/parenting programs when they were six-year olds. We consider various channels of influence, inclu- ding occupational choice, fertility effects, and higher earnings stemming from the additional educational Source: Staff elaboration (See Technical Endnote 11). attainment. Only considering these channels, we find significant impacts with, for that generation, incomes Reflections on the Public Policy Debate being almost 8 percent higher, the poverty rate 11 percent lower, and the female labor force participation 32. To summarize, our report has several broader rate more than 9 percent higher. Such simulations are findings. Inequality of opportunity is important in partial as they do not take into account synergies with explaining both the distribution of wealth and education other ECD programs (like early cognitive development), outcomes in Turkey today. Such inequity, we find, 5 Both models concentrate on assessing direct impacts only and hence do not assess overall general effects which would have to include many more behavioral functions. Turkey: Expanding Opportunities for the Next Generation xii travels through generations with child poverty and complemented by an integrative and inclusive policy stunting closely mapping into the socio-economic for the most disadvantaged children. background of grandparents. Taking a closer look at children in such low opportunity settings, we discover 35. Third, international evaluations show that the clear disadvantages with regards to access/inputs into most effective way to reduce the influence of their development process which show them lagging circumstances on opportunities is to provide effective behind early on in physical and cognitive development. support to the most disadvantaged children first. The concept of equality of opportunities employed in Turning to policy, Turkey spends relatively little on its this Report goes beyond creating equal access – it youngest generation today, with low coverage rates implies that the most disadvantaged are reached first across most dimensions of early childhood development, and more intensively than less disadvantaged children especially of the disadvantaged, being low. These so that they can improve their life chances. If this might, though, be the highest return investment into the concept resonates in Turkey, a strategy for roll-out country’s future, as several of our simulations show. would then need to set targets detailing, how and how many of the most disadvantaged children can be reached. 33. The report ends with a number of reflections for Currently, vital child services – from nutritional support, the public social policy debate. First, to improve health attention as well cognitive stimulation – largely equity in society, opportunities for disadvantaged benefit less disadvantaged children, those that (judged children would need to be expanded. This will by their own circumstances) are more likely to succeed necessitate reviewing the current functioning and in life at any rate. While a discussion on how financing of the Turkish welfare state. Current public, ‘disadvantage’ can be defined will be necessary, this non-contributory social expenditures reach children in report suggests that two factors alone, child poverty their early years to a limited extent. To create fiscal and parental education, are core determinants of space that would allow programs for disadvantaged opportunities. children to expand, the financing and societal (tax financed) transfers to old age insurance would need to 36. Lastly, civil society, community, and private initiatives, will need to complement public efforts in be re-examined. expanding the supply of services for the most disadvantaged children. Turkey’s innovative and 34. Second, Turkey’s informal safety nets, as strong inspirational experiences of ECD service delivery as they might be, do not seem to have been able to through non-governmental channels is looked at across offset the disadvantages of children born into specific the globe with much admiration. In parallel, overall circumstances, especially for girls. Turkey has a coverage of such delivery is very low. A social consensus traditional and strong communal and family solidarity to support the most disadvantaged children would require which is often described as being one of the main pillars a compact between private, public, civic, and community of the societal functioning. But given the strong inter- actors to create an appropriate enabling environment generational transmission of inequity observed here, for the availability of high quality and integrative support such communal and traditional ties would need to be for disadvantaged children. Turkey: Expanding Opportunities for the Next Generation 1 1. Introduction: Life Chances 1. This Report examines life chances. Life chances Ayþe’s chances of being fully immunized when she for today’s Turkish people, most importantly the future celebrates her first birthday are only a third of his generation, today’s children. The results presented in chances (24 against 72 percent). And she is ten times this Report show that life chances differ in important more likely to suffer from stunting, a sign of chronic dimensions today – and that Turkey could immensely malnutrition in which children show growth deficiencies improve its human and economic development potential (about 30 against 3 percent). Similarly, the two children’s by maximizing such opportunities. education trajectories are likely to differ starkly as well: Ayþe has a small chance (about one in five) of completing 2. To illustrate, imagine a girl, Ayþe, and a boy, secondary school while Mehmet, with a very high Mehmet, to be born on the same day but in two very likelihood, will complete secondary school successfully 6 different environments: The girl’s parents are poor, and move on to university. At age fifteen, learning, commanding around 80 TL monthly for every member reading and comprehension differences will most likely of the family to meet the costs of food, clothing, be fundamental. Later in life, such diverging education schooling, transport and other life essentials. She is achievements will lead to widely different opportunities born in a remote village in Eastern Turkey, located of finding good jobs and earning incomes that would between Erzurum and Aðrý. Imagine both of Ayþe’s allow Ayþe and Mehmet themselves to offer good life parents working in animal husbandry with her eldest chances to their own children. sister looking after the little baby and her siblings. Her small house, made of bricks, holds few belongings and 4. This Report sets out to explore a number of questions the family does not own a television, washing machine about diverging life chances in Turkey. Are opportunities or a car. Imagine Mehmet, on the other hand, growing in Turkey shaped by such characteristics of the up in a well-off neighborhood in Izmir and a quite environment mentioned above – like birth place, wealthy household able to afford monthly expenditures education and wealth of one’s parents, or even language of more than 1,000 TL for every person in the household. spoken in one’s childhood? If so, which characteristics Mehmet’s parents, who have completed university are the most important ones? And are there other factors degrees, are both working in professional jobs. He will of importance in understanding life chances – those grow up in a large apartment which holds all assets that that reach further back like the education of Ayþe’s and contemporary urban life offers. Mehmet’s grandparents? Lastly, how many children are affected by such low opportunities – and how can 3. Comparing the life trajectories of these two, society support them to reach their own, full potential? imaginary children until age fifteen is telling. Chances begin to differ already at birth – or even before: Based 5. We aim to shed some light on these questions in on recent patterns observed, the chance of Ayþe’s mother the following way. As background, we shortly review to receive regular antenatal care during pregnancy is trends in poverty and social development in Turkey only about 45 percent while Mehmet’s mother is almost over the past years, contrasting such developments with certain to have followed such check-ups. The girl’s a recent representative opinion survey that provides an risk of being born with a low birth weight is one in four insight into the aspirations of Turkish society. Section – or more than four times higher than that of Mehmet. Three introduces the equality of opportunity concept And such differences continue through to school age: and applies it to examine wealth and education outcomes. 6 The environment and chances of the two children are derived from examination of the 2004 Demographic and Health Survey with monetary values inflated to 2009 prices. Monetary values are illustrations only and the examples fall in the classifications of intergenerational opportunity groups introduced in this Report. See Technical Endnote 1. Turkey: Expanding Opportunities for the Next Generation 2 We turn to examine how other, intergenerational, factors considers as the most effective policies to support are linked to the well-being of children in Turkey today disadvantaged children to reach their development – most prominently the education of the maternal potential – and for which the Government of Turkey grandparents. Section Four explores child wellbeing has set itself high aims: integrated, holistic policies and and child risk factors while, lastly, Section Five programs that support disadvantaged children and their concentrates on what this Report, like many others, families from the earliest moment in their lives. Graph 1: Under Five Mortality Rate Over Time (1960-2006) Source data: World Development Indicators (The data for 1965 is extrapolated from existing data). 2. Equity Aspirations 6. From the Marmara earthquake and the 2001 financial in secondary school, often characterized as the Achilles crisis to the onset of the global economic slowdown in heel of human development in the country, climbed 8 mid-2008, Turkey recorded major social and human steeply from 51 percent (2002) to 59 percent (2008). development progress. Under-five mortality rates Similarly, poverty had decreased from 27 percent in continued their remarkable, decade-long decline, 2002 to 19 percent in 2007 – and further decreases have reaching 23.9 (per thousand live births) in 2008.7 As likely materialized until mid-2008. Such poverty Graph 1 shows, Turkey achieved such sustained reduction was achieved not only through the strong reduction even during the times of economic decline growth performance of the economy but also due to a and is now outperforming countries in the Middle East marked reduction in inequality in society: between 2003 and North Africa that for a long time held similar levels and 2006, consumption inequality declined by more 9 of the under-five mortality rate. Net enrolment rates than ten percent. 7 Preliminary results of Turkey Demographic and Health Survey (2008), Hacettepe University. 8 Ministry of National Education, Statistical Yearbook, 2008-2009. 9 Aran et. al. (2009), one of the working papers jointly produced by the Government and the World Bank on social policy, has a detailed analysis of the contributions of changes in inequality and average consumption to poverty reduction in Turkey between 2003 and 2006. Inequality, as measured by the adult equivalence adjusted consumption Gini coefficient, declined from 34 percent to 31 percent. Turkey: Expanding Opportunities for the Next Generation 3 Graph 2: Should the Gap Between the Rich and the Poor Graph 3: What is the Main Reason why there are some in Turkey be Reduced? people in need in Turkey today? Source: Life in Transition Survey (2006) Source: Life in Transition Survey (2006) 7. Such improvements, welcome as they certainly are, it – on the contrary. The opinion survey also finds that still lag the aspirations of the Turkish people, as three quarters of Turkish people feel that success is due eloquently documented in a 2006 opinion survey. The to effort, hard work, intelligence or skills. 10 Life in Transition Survey was conducted in Turkey and twenty-seven countries of Eastern Europe and 9. Some divergences in social indicators, often identified Central Asia and asked a nationally representative group with potential underlying inequities in societies, are of adults about their opinions on living standards, indeed significant in Turkey today. While net secondary poverty and inequality, trust in state institutions, and school enrolment rates have increased, as observed above, attitudes to the market economy and public services. the gap between girls’ and boys’ completion rates remains 11 A full 85 percent of the adult population expressed that significant, at 9 percentage points in 2006. As the gap between the rich and the poor in society should prominently pointed out by the Education Reform be reduced – the single highest proportion in any country Initiative (2009) and also emphasized later in this Report, in which the survey was fielded (Graph 2). And an girls’ learning achievement is at least equal to that of even higher proportion, 92 percent, believes that the boys – so that different school attainment rates are often state should be involved in reducing this very gap interpreted as representing underlying inequities. between the rich and the poor. 10. Similarly, the female labor force participation rate 8. The same survey also solicits opinions about justice remains, with 25.5 percent in 2008, at very low levels. and fate. Two out of three Turkish adults responded Graph 4 visualizes that the low level of the female that the main reason why poverty existed in the country participation rate as well as its large gap to the male was ‘injustice in society’ (Graph 3) – if we consider rate stand in stark contrast to emerging market and that ‘luck’ and ‘inevitable part of modern life’ also refer OECD countries. A companion report from the World 12 to factors beyond the control of the individual, then a Bank shows that educational attainment is strongly full three-quarters of the Turkish population feel that correlated with female labor force participation rates – the poor should not be held responsible for their so that the observed participation gap could, at least in condition. Such opinion does not imply, however, that part, be due to inequities materializing through the those who did obtain wealth and standing do not deserve education system and beyond. 10 For descriptions and results of the surveys, see European Bank for Reconstruction and Development (2007) and Zaidi et al. (2009); Ferreira et al. (2009) discusses the Turkey results. 11 World Development Indicators (2009). School completion rates, by gender, are only provided in the World Development Indicators until 2006. More recent data is available only for net enrollment (and not completion) rates for secondary school which are reported by MONE at 60.6 percent for males and 56.3 percent for females as of 2008. 12 World Bank (2009b). Turkey: Expanding Opportunities for the Next Generation 4 Graph 4: Labor Force Participation Rates, by Gender Source: World Development Indicators, 2009 3. The State of Equality of Opportunity in Turkey13 In this section, we assess equality of opportunity that the social environment (family, household, quantitatively in Turkey. We define equity as a state in community, education and health system, etc.) is which circumstances at birth – which are not dependent supportive of her or his development needs and provide on one’s own effort or luck – do not influence life the best possible basis for the child to choose her or his chances. Examining both the distribution of wealth self-determined path in life. and education outcomes in Turkey today, we find that inequalities of opportunity are important in society 12. Transforming such opportunities into actual today, explaining about one third of the overall variation achievements depends on one’s own drive, effort and, in such outcomes between households. We discern a at times, luck. But the opportunities themselves might number of circumstances relating to birth area and be determined by factors that no one can do anything socio-economic background which are powerful about: in which family one is born and where; the determinants of advantage and disadvantage in society educational background of one’s parents and their today. And we also find that lacking opportunities wealth; the language spoken in the family; or one’s travel between generations and that the chances of 14 today’s children are closely related to the socio-economic gender. Such factors, or circumstances, as John Roemer background of their grandparents. calls them, are all independent of one’s own choices and effort. 11. What concept does ‘equality of opportunity’ stand for? First of all, and following the concept laid out in 13. A society offering equal opportunities to its citizens he World Development Report 2006, by opportunities, would then be one in which all those circumstances, we refer to possibilities of people to succeed in life – such as the socio-economic family background or birth to lead healthy, long lives, free of material or social place, matter little or nothing in determining life chances. deprivation. For a child, for example, this would mean One in which a girl born in a remote, poor, area between 13 This section draws on Ferreira et al. (2009) and Ferreira and Gignoux (2009). 14 Roemer (1998). Turkey: Expanding Opportunities for the Next Generation 5 Erzurum and Aðrý to parents without much formal Inequality of Economic Opportunity education would have the same chances to succeed in life as a boy born in the center of Izmir to wealthy -and 17. For the analysis of inequality of economic educated parents. opportunity, we are able to distinguish six circumstances, factors that are outside the scope of influence of an 14. The equity concept used here places much emphasis individual. Both the number and choice of these on opportunities rather than on outcomes per se. Equality circumstances is determined by availability in our data of outcomes would imply that reading scores at age source, the Turkish Demographic and Health Survey fifteen of all children should be equally high, that wealth (TDHS). The TDHS collected such variables for all be distributed homogenously, or that life expectancy ever married-women between 15 and 49 years of age. should be the same for everyone in society. But such Importantly, more than 80 percent of all households in outcomes also depend, to a significant extent on one’s Turkey have at least one ever-married woman in their own efforts, disposition or luck. For example, lifestyle midst so that the survey covers a very large part of the 15 decisions (e.g., whether to smoke) will be an important population. The circumstances we consider include determinant for individual life expectancy. Also, the woman’s place of birth (rural/urban areas, as well rewarding effort (to learn, study and work) is essential as region), their father’s and mother’s education, the in societies for innovation and advancement. number of siblings in the family, as well as the language spoken in the household when she was growing up. 15. The concept of equality of opportunity does, however, imply more than providing universal access 18. Relationships between these circumstances at birth to basic public services like education. As we will see and outcomes in later life can be both direct and complex later in the Report, differences in child well-being at the same time. While we aim to assess the aggregate emerge early in Turkish society today – differences that influence of such circumstances, we do not aspire to have nothing to do with individual effort or luck. For understand specifically how each of them influences example, according to the 2004 Demographic and Health life chances. For example, birth place can exert its Survey, twelve percent of children below the age of importance via many channels: via the quantity and five were stunted – a measure for chronic malnutrition quality of available health services or schooling; via – which has been proven to have a strong negative access to public services such as sanitation and clean effect on learning abilities. This form of malnutrition water (reducing health hazards) or electricity (influencing is closely related to the poverty and socio-economic information connectivity as well as allowing reading circumstances in the households in which children grow and communication during darkness); via connectivity up. Hence, even if school access was universal and to markets which could bring employment and income- education quality high across the country, it is unlikely earning possibilities; or via the availability of strong that those children stunted when infants would be able family and community networks that provide support to learn equally well than children that were well- and encouragement throughout child- and adulthood. nourished when young. Equality of opportunity would Similarly, a birthplace might also be associated with imply an active effort to reach disadvantaged groups in positive or negative discrimination later on in life if a society and to offset such disadvantages early on. geographical area is identified by society-at-large with 16 specific attributes. 16. As evidence that opportunities are not equally distributed in Turkey, we find that circumstances are 19. We are able to include in our analysis only some important in determining success in Turkey today. In of the many circumstances that will influence economic the remainder of this Section, we present quantitative outcomes – therefore, the estimates presented can safely estimates of inequality of opportunity for two outcome be judged as conservative, denoting only the minimum dimensions: economic and educational achievements. inequality of opportunity present in Turkish society 15 See Technical Endnote 1 which presents which household types are included in the sample and which ones are not captured. 16 See Wietzke (2009) for a recent review of this literature, specifically as to how spatial variables can impact on poverty and human development. 6 Turkey: Expanding Opportunities for the Next Generation today. For example, we could not include any information housing conditions (size, materials), as well as access 17 pertaining to the husband of the ever-married women to amenities (public service supply). The relationships (birthplace, his parents’ education, etc.) nor regarding between wealth and the circumstances are depicted in household income or wealth into which husband and Graph 5. Each panel of the graph shows how household wife were born. All these other circumstances will also wealth – ranging from the least-wealthy on the left to the wealthiest on the right – and a circumstance are exert an influence on opportunities – hence their inclusion associated. A strongly downward sloping curve signals would lead to an increase in the measured inequality of a strong association. For example, panel A shows that opportunity in Turkey. 85 percent of women in the least-wealthy households were born in rural areas while this holds only for about 20. In Turkey, wealth and the measured circumstances 20 percent of women living in today’s wealthiest are closely related. We measure wealth by a household’s households. Note that this panel depicts the birth place ownership of durable goods (ranging from bicycles to of the ever-married woman in the household and not tractors and from refrigerators to air conditioners), where family members currently live. Graph 5: Wealth Distribution and Circumstances, 2004 Source: Staff elaborations based on Ferreira et al (2009) 17 Ferreira et al (2009) build on the methodology developed by Filmer and Pritchett (2001) to define the wealth variable. Turkey: Expanding Opportunities for the Next Generation 7 21. Other circumstances show equally strong asso- 23. Examining the relative importance of the different ciations with household wealth. This holds importantly circumstances, we find that area and parental education so for the education of both parents (panel C and D). stand out. Looking at the decomposition of this share 21 Comparing the shapes of these two panels also reveals (which has to be interpreted with caution ), we find a remarkably different distribution of educational that birth area (rural or urban) appears as the most attainment of the elder generation by gender in Turkey: important factor impacting on opportunities. There are roughly, the ‘parents generation’ (or the grandparents many venues through which area and region can of today’s children) depicted here was born between influence today’s distribution of wealth in Turkey – and 1940 and 1970. The share of women in that generation access, reliability as well as quality of public services not holding a primary school diploma was significantly (including education services) can be one of them as 18 higher than the corresponding share for men. are connectivity to markets and access to information. Following birth area, parental education – which is 22. Assessing the importance of all such circumstances closely related with socio-economic status – emerges together, we find that at least one third of the wealth as a similarly important circumstance factor. These disparity in Turkey is due to inequalities of opportunity three factors alone account for two thirds of inequality (Graph 6). Again, since we were not able to capture of opportunity related to the wealth distribution in the all circumstances, this is a conservative estimate with country. Mother tongue, number of siblings and birth the inequality of opportunity share most likely being region follow, albeit with lesser significance. higher. The techniques applied assess the share of wealth inequality which is due to ‘circumstances’ and, 24. An opportunity profile for Turkish society can depict 19 hence, inequalities of opportunity. We find that birth how prevalent different exogenous circumstances at circumstances have a powerful and strong influence on birth are. The opportunity profile is constructed by life trajectories. Limited analyses for other countries using all circumstances and defining population groups exist that would allow us to compare Turkey – e.g., all households whose ever married woman was internationally – as a rough indication, the country born in the rural, Western area of the country to parents appears to rank towards the more moderate end of Latin with completed secondary school degrees and where American countries; a continent that has long been Turkish was spoken at home. Such groups can then be 20 highlighted for the persistence of inequalities. aggregated so as to include the most advantaged ten percent and the least advantaged ten percent of the population using their observed household wealth. The Graph 6: Inequalities of Wealth Opportunity in Turkey ‘least advantaged’ ten percent would then encapsulate the population with the least opportunities to acquiring 22 wealth, based on exogenous circumstances at birth. 25. We find a high concentration of circumstances for the opportunity deciles, indicating that a core group of households is cumulating circumstances that are associated with disadvantage. As depicted in Graph 7, for the most advantaged (least advantaged) decile, we find that 99 (3) percent of ever-married women were born in urban areas, and about two-thirds (4 percent) Source: Ferreira et al (2009). in the Western part of the country. A very high share, 18 We use the previous definition of primary schooling (five years) for this calculation. 19 See Technical Endnote 2. 20 See Ferreira et al. (2009) and Paes de Barros et al. (2009). 21 See Technical Endnote 3 and Ferraira et al (2009) for details on calculations and methodology. 22 See Technical Endnote 4. Turkey: Expanding Opportunities for the Next Generation 8 94 percent (12) have educated mothers with at least a child and 85 (4) percent had three or less siblings. primary school diploma and virtually all, 99.7 (19) Hence, while we found several exogenous circumstances percent had educated fathers with at least a primary to explain a large share of the inequality of opportunities diploma as well. Further, more than 98 percent (9) in Turkey, the overlap with other circumstances percent had Turkish as a mother tongue when as a young (including region) is high. Graph 7: Opportunity Profile for Least and Most Advantaged Population Groups, 2004 (percent) Note: Circumstances defined for ever-married women. Source: Ferreira et al (2009). (See Technical Endnote 4) Inequality of Opportunity for Educational to underlying inequality of opportunities. When all Achievement circumstances are considered together and controlled for, family background emerges as the dominant source 26. We now turn to examine how exogenous circum- of such inequality for education achievement: Three stances at birth influence a non-economic outcome, quarters to four fifths of the inequality of opportunity 23 namely educational achievement. For that purpose, can be accounted for by socio-economic variables of the inequality of opportunity concept can be applied to the students’ families, composed of parental education, data from the 2006 Program for International Student asset ownership and father’s occupation. Assessment (PISA) which recorded standardized tests for reading, mathematics and science. It was applied 28. Spatial variables, both area and region, contribute to a large sample of 15-year old pupils in Turkey and about one fifth to the opportunity inequality. Schools recorded, in parallel, information about the gender of located in the East or in rural areas are associated with the student, area and region of the school, information lower test scores, even when other circumstances, such on the family background (father’s occupation, parental as family background, are taken into account. But these education), as well as a number of asset (wealth) spatial variables account for no more than one fifth of variables. the lower-bound overall opportunity share of inequality in achievement in Turkey. For reading and science, the 27. Similar to the results regarding the distribution of rural-urban divide is more important than the broad wealth, we find that between a quarter and a third of regional location – an analogy to the results regarding overall educational achievement variation can be traced the determinants of inequality of economic opportunity. 23 See Ferreira and Gignoux (2009). Turkey: Expanding Opportunities for the Next Generation 9 29. While gender emerges as a key associate of inequality. Socio-economic family status, including enrollment, as we will report later on, it is not a significant education, retained its paramount importance in the determinant of achievement. Girls performed better than education achievement analysis as well. These results boys in reading (significantly so) and somewhat worse also showed that spatial variables – both area and region 24 in mathematics in 2006 according to the PISA results. – remained important, albeit to a lesser extent than Overall, from the PISA results, if girls stay in school, socio-economic status. Based on these findings and there is no evidence that girls perform worse than boys. the patterns observed when analyzing the opportunity profiles, we define those children as in the Low 30. These findings support the recent and prominent Intergenerational Opportunity Group (LINOG) if their education equity analysis conducted by the Education mothers were born in the Eastern, rural, area with both Reform Initiative (ERI, 2009). The ERI report finds their maternal grandmother and grandfather not having that students are being separated into different quality completed primary education. Conversely, the High schools at the secondary level based on their socio- Intergenerational Opportunity Group (HINOG) is economic status – and that the basis for such separation composed of children whose mother was born in an stems from lacking access to pre-school education and urban area in the Western or Central region and whose quality primary education for disadvantaged children. maternal grandparents both completed at least primary Given that secondary school types (Anatolian high school (see Table 1). school, general high school, vocational high schools, multi-track high schools) are a core determinant of Table 1: Intergenerational Opportunity Groups of Children university access – and higher earnings in the future – ERI concludes that the education system in its current Birth place Education of maternal form restricts social mobility and has not realized equal of Mother grandparents opportunities for all children. Low Intergenerational Less than Opportunity Group Rural East primary (LINOG) completed Grandparents and Grandchildren High Intergenerational At least Opportunity Group Urban primary 31. We now turn to explore whether the observed (HINOG) West/Center completed inequalities of opportunity influence also today’s children Source: Staff elaboration. – hence to what degree they travel through generations. Comparing Turkish society today with the social and economic life of four or five decades ago reveals that tremendous transformations have taken place – regarding 33. Significant shifts have taken place between the urban and rural life and livelihoods alike. But such grandparents’ and children’s generation included in the transformations notwithstanding, this Report finds that low intergenerational opportunity group. The most the socio-economic status of grandparents – measured important of these relates to current location. While by the education they attained roughly forty or fifty all grandmothers gave birth to their daughters in the years ago – retains a powerful link to the wellbeing and rural, Eastern, part of the country in the LINOG, the chances of their grandchildren today. 25 current domicile of the grandchildren’s generation is much more disperse – an effect of the strong migration 32. For illustration purposes, we distinguish two groups that took place over the past decades (Table 2). Only to trace the intergenerational transmission of opportunities. about 50 percent of children belonging to the LINOG The analysis above had highlighted the prominence of live in rural areas of Turkey’s eastern region today but area (rural/urban) and parental education as circumstan- an additional 30 percent, having moved to urban areas, ces that explain the largest opportunity share of wealth remain in the Eastern region. 24 The better reading outcomes for girls could be influenced by the high drop-out rate for girls at secondary levels. If those girls with worse test-scores were to drop-out over proportionally, the average scores of the remaining cohort increases. 25 Given that the TDHS data only includes circumstance variable for ever-married women, we are only able to examine the relationship between maternal grandparents and their grandchildren. Turkey: Expanding Opportunities for the Next Generation 10 Table 2: Intergenerational Opportunity Groups and Current the LINOG was with 78 percent eighteen times as high Residence of Children, percent as for the HINOG group in which child poverty barely East East West/Central West/Central reached 4.4 percent. Examining poverty risks for Total Rural Urban Urban Rural different sub-groups of the LINOG, we find that the LINOG 50 30 18 2 100 poverty rate of those families that migrated to urban areas in Western and Central Turkey is, with 40 percent, HINOG 3 0 90 7 100 lower than for those that remained (89 percent in rural Source: Staff Elaboration based on Turkey DHS (2004) Eastern areas, 73 percent in urban Eastern areas) but it remains almost three times as high than the average poverty rate in urban non-Eastern areas. Hence, regional migration is associated with reduced but not vanishing 34. Examining educational achievement of the children’s disadvantage. Opportunities might be improving but parents in the two groups, we do find a significant share not, as we had seen in Section Two, at the speed that of reproduction of educational attainment across the Turkish society aspires to. generations. Table 3 shows that educational advance in the LINOG has been limited, especially for women: Graph 8: Height for Age Measures for LINOG and HINOG two-thirds of today’s children in the LINOG have a mother without a primary school degree – these women havenot advanced in their educational achievement compared to their own mothers. The children’s fathers did record a significant share of obtaining a primary school degree. Still, more than 80 percent did not achieve to complete secondary school. Table 3: Intergenerational Opportunity Groups and educational attainment of grandparents and parents, percent Less than primary/ Primary Secondary Higher Total illiterate education Note: Staff estimation based on TDHS 2004 LINOG 68 31 1 0 100 child’s mother child’s father 23 61 14 2 100 37. The results for stunting are equally pronounced. Out HINOG of one hundred children below the age of five, 27 in the child’s mother 0 31 51 18 100 LINOG showed a low height for age, an indicator that child’s father 0 20 55 24 100 indicates the long-term effects of malnutrition – while in Source: Staff Elaboration based on DHS (2004) the HINOG only four out of one hundred children showed such signs. Graph 8 shows how divergent the growth paths of children in the two different groups are. The 35. The two groups are of similar size. The LINOG vertical axis indicates the density of the distribution – how comprises a little more than 14 percent of children, the many children reach a certain height. The horizontal axis HINOG close to 16 percent in 2004, the base year for our is an indicator of the actual height they achieve – with the analysis. As we had seen when examining the opportunity ‘0’ indicating the expected, normal, height for their age. profile, circumstances are highly correlated. Of the children We see that the two curves diverge strongly, with the in the LINOG, about 11 percent have a mother whose HINOG curve peaking much more to the right, in the scale mother tongue was Turkish; in the HINOG, this is the that signals normal growth performance. For the LINOG, case for 98 percent of the children. on the other hand, we find the curve shifted much more to the left with many children achieving a significantly 36. We find that the inter-generational opportunity worse than normal growth performance. groups are very closely aligned with the well-being of present day children. The most direct result is the close 38. We find that lack of opportunity travels through relationship between child poverty, in 2004, and the generations and significantly influences the development intergenerational opportunity groups: child poverty in potential of disadvantaged children in Turkey today. Turkey: Expanding Opportunities for the Next Generation 11 Graph 9 visualizes the contrast between the inter- 39. In addition, we find that girls appear to be impacted generational opportunity groups with respect to the risks stronger by such inter-generational transmission of of poverty and stunting. Stunting is, according to cross- opportunities. Within the LINOG group, young girls are country research, among several leading acute risk factors significantly more likely to show early signs ofmalnutrition that signal children not reaching their full cognitive than boys: their stunting rate is with more than 30 percent 26 development potential. As the Graph shows, a child about a third higher than their malesiblings (at 23 percent). born in the LINOG group in 2004 (red bars) was 1.5 Given that stunting is an early sign of acute risk for 27 times more likely to be stunted than the average stunting children, this result is particularly worrying. Child level in Turkey. development trajectories continue to differ in education: Graph 9: Relative Risk of Child Poverty and Stunting, by in the LINOG the likelihood of girls in the ages of 7 to Intergenerational Opportunity Group, 2004 15 being enrolled was 68 percent – compared to around almost 90 percent for boys. 40. Hence, breaking the intergenerational transmission of inequity would have to place particular emphasis on supporting disadvantaged girls in Turkish society today. Given their socio-economic background and other circumstances at birth, they are more likely to lose in the combination of low opportunities and poverty. Providing focused support to disadvantaged girls so that they can reach their own development potential, including education, would be instrumental in improving the socio- Note: Risk relative to average incidence of all children age 5 and below economic circumstances in which their own children are Source: Staff calculations based on Turkey Demographic and Health Survey (2003/2004) born in the future. 4. Child Development and Child Risks Our finding that the intergenerational transmission of Child Poverty inequity is powerfully affecting Turkey’s youngest generation today points to the need to understand how children’s 41. Given the very close mapping of the intergenerational opportunities develop from a young age, and indeed opportunity profile to child poverty, one important whether there are policy interventions that can contribute indicator of life chances for today’s children is their to reduce the impact of exogenous circumstances on life poverty status. Poverty indeed is a circumstance for chances. This section then takes a more in-depth look at children in early ages given that it is defined at the child well-being in Turkey. We start with a closer look at household level and children do not contribute to the child poverty given the very stark divergence between the income or asset envelope of households. Cross-country likelihood of children growing up in poverty between research has shown that poverty is one of the most opportunity groups. We then map child development by important factors linked to development delay of poverty and opportunity groups along different child children. It is often associated with a number of variables development dimensions. Lastly, international that directly impact on child well-being, ranging from benchmarking shows that Turkey’s current ambitious child inadequate food intake, poor sanitation and hygiene development targets are well grounded given that the facilities that can lead to infections and growth country falls short in relation to comparator countries in retardation. Similarly, poverty is often correlated with several dimensions. parental, especially maternal, stress as well as lower 26 Grantham-McGretor et al. (2007) and Walker et al (2007). 27 In multivariate regression analysis, we also find that when linking child access and outcomes to circumstance variables, girls have a significantly lower probability to obtain immunization and medical treatment than boys. Turkey: Expanding Opportunities for the Next Generation 12 Graph 10: Relative Risk of Poverty, by age group 2003 and 2006 Graph 11: Distribution of Estimated Numbers of Additional Poor, by Age Group, 2009-2010 Source: Staff elaborations based on Aran et al (2009). Source: Staff estimations. (See Technical Endnote 5). education levels of the parents and less cognitive economic slowdown Turkey is witnessing today. Based 28 stimulation. on a simulation model that links the overall economic slowdown to the reduction in employment in different 42. In 2006, child poverty was higher than for any other sectors of the economy, the potential rise in poverty could population age group in Turkey. Both younger (zero be largely made up of children in the age group zero to to five years) and older (six to fourteen years) children fourteen. As visualized in Graph 11, we estimate that had the highest poverty rates among all age groups with almost a third of the additional poor could be made up more than one in four children being poor. Forty one of children below the age of 14. 30 percent, or over five million, of all poor people in Turkey in 2006 were children below fourteen years of Child Poverty and Opportunity Groups age. About 1.9 million of them were infants and young 29 children under five. 45. Findings from early childhood development research stress that it is the multi-dimensional lack of opportunities 43. While overall poverty rates had decreased between that put children at highest risk of not reaching their 31 2003 and 2006, children benefited least from such development potential. Lacking opportunities can stem improvements in welfare. Graph 10 illustrates the from many different sources – including those examined relative poverty rate of children – hence measuring the above (i.e., poverty, parental education, gender, spatial risk of children to be in poverty compared to all other characteristics) and factors that are more difficult to age groups in Turkish society. Such relative risk for capture (i.e., functionality of the family or extended children zero to 5 increased slightly and the risk of family, neighborhood support). Childhood development those aged 6 to 14 rose quite strongly between 2003 will be especially hampered if children face many lacking and 2006. opportunities at the same time. 44. According to estimations carried out for this Report, 46. Using such accumulation of lacking opportunities, children are also the population group that is likely to we examine child development in Turkey in relation to be at highest risk of falling into poverty due to the lacking opportunities along a simple measurement 28 Grantham-McGregor et al. (2007), p.62. 29 Calculations using Aran et al (2009), p. 35. 30 Based on Hentschel and Aran (2009). See Technical Endnote 5. 31 The ecological model of child development emphasizes the large number of environmental influences on children and rather than singling out the partial impact of various risks finds that it is the cumulative combination of risks that increase the probability of children not reaching their development potential (Shonkoff and Meisels 2000). Turkey: Expanding Opportunities for the Next Generation 13 framework presented in Graph 12. First, we look at a somewhat later in life – educational attendance diverges number of child access indicators of health services according to opportunity groups and poverty. These (e.g., birth attendance), administrative functions (e.g. three dimensions form a measurement framework that birth certificate) and inputs for cognitive stimulation. allows tracing child development albeit we will have to Second, we look at the presence of acute risk factors be cautious about assigning causality as many other factors for development – hence early signs of development will influence child development outputs and outcomes. gaps emerging for different groups of children. These With respect to measures of lack of opportunity, we retain include low-birth weight, stunting and cognitive our intergenerational opportunity group classification 32 development retardation. Lastly, we measure how – (LINOG an HINOG) and add to this child poverty status. Graph 12: Child Development Measurement Framework Source: Staff elaboration based on World Bank (2009b). 47. We find core access indicators to differ strongly by the children are born, with two thirds of poor mothers in opportunity group and child poverty status in Turkey. Table low opportunity settings not receiving a minimum of 4 shows, for 2004, the different access indicators for poor antenatal care during their pregnancies. Similarly, more children and those children in the LINOG and HINOG. than 92 percent of poor children in the LINOG group do Across the board, poor children in low opportunity settings not command food with sufficient iodine supply and four show remarkably worse access indicators to basic health fifths do not receive the full set of six recommended functions than other groups – and this starts already before immunizations before they complete one year of age. Table 4: Child Development Access and Outcome Indicators by Opportunity Status, 2004, percent Child Access (input) Indicators, percent Child Development Outcomes, percent Incomplete No Unattended Insufficient Immunization No Low antenatal care Births Iodine Supply (first year) birth certificate Birth Weight Stunting All children 24 21 50 48 16 11 12 -- poor 50 45 79 68 28 18 23 -- non-poor 14 8 33 36 9 9 6 LINOG 54 56 81 75 28 22 27 -- poor 62 65 92 80 30 24 30 -- non-poor 21 20 35 58 20 20 15 HINOG 4 4 20 27 6 6 4 -- poor 25 18 55 55 33 na 14 -- non-poor 3 3 19 26 5 6 4 Source: Staff calculations based on TDHS (2004). (See Technical Endnote 6.) 32 Paes de Barros et al (2009) introduce a new measurement tool which can be applied to individual access rates to basic services, the Human Opportunity Index. The index captures both access levels and their distribution in the population, giving more weight to poorer households as these are likely the ones with lacking opportunities. Further work in Turkey could build on the Human Opportunity Index concept and measure both regional opportunities as well as changes over time. Turkey: Expanding Opportunities for the Next Generation 14 48. The cumulative nature of the lack of opportunities 49. New data from Koç University show that inputs becomes equally visible when examining these access for cognitive development of children differ strongly 33 indicators. Differences within the LINOG and HINOG by socio-economic strata (SES). While we cannot (between poor and non-poor children) as well as between present this data along the opportunity group/poverty poor children (across the groups) show some marked dimension as in Table 4, we have observed that socio- differences. For example, poor children in the HINOG economic strata and opportunities for children are closely – representing only 5 percent of children within this linked. Graph 13 exemplifies that crucial inputs into group – are twice as likely to be issued a birth certificate the cognitive learning process for children – e.g., as poor children in the LINOG group. Similarly, non- language stimulation, availability of learning materials poor children in the HINOG are ten times more likely – differ strongly according to socio-economic strata for to be born in a medical facility than poor children in the 3-4 year olds. the same intergenerational opportunity group. 50. Such stimulation has been proven as a core Graph 13: Cognitive Development Inputs at Household Level for determining factor for cognitive development and later 36-47 months old children in Turkey educational achievement – and this relationship can be 34 observed in Turkey as well. The ongoing research study by Koç University also collects cognitive test scores for children between the ages of 3 and 4. Test results show an early diversion according to socio- economic status of the families in which the children grow up: Graph 14 provides language comprehension (Tifaldi) and short term memory (Corsi) test results and visualizes the stark difference in cognitive development outcomes between children from different socio- 35 economic backgrounds. Source: Data from the Study of Early Childhood Developmental Ecologies in Turkey, 2008, 36-47 months old children and their families. 51. Moving to the outcome dimension in the measure- Graph 14: Cognitive Development Scores of 36-47 months old ment framework of Graph 12, we find that they signal children in Turkey acute risk factors for children in low opportunity settings. Low birth weight affects a quarter of poor children in the LINOG (Table 4). Low birth weight indicates constraints in fetal nutrition during a crucial period for brain development – and it is largely due to poor maternal 36 nutrition and infections. Similarly, high stunting rates are concentrated in poor children of the LINOG group as well, albeit they are also significant for poor children in the HINOG group – indicating that acute risk factors can also develop in poor households with a better Source: Data from the Study of Early Childhood Developmental Ecologies in Turkey, 2008, 36-47 month old children and their families. educational background of the parents. 33 Study on Early Childhood Development Ecologies in Turkey, Koç University. The study collects longitudinal data for the children. 34 Shonkoff and Phillips (2000). 35 The Corsi test is an internationally standardized short-term memory test. The Tifaldi language test is designed specifically for Turkish language comprehension. 36 Walker et al (2007), p. 145. Turkey: Expanding Opportunities for the Next Generation 15 Graph 15: Enrollment Age Profiles by Opportunity Group and Gender, (2004) Source: Staff elaborations using TDHS 2004. International Comparisons 52. Lastly, moving to medium-term impacts, we find 53. Internationally, Turkey performs less well than that lower access rates and the emergence of acute risk comparator countries with respect to core child development factors in early childhood go hand in hand with sign- indicators. Benchmarking of indicators, such as ificantly diverging enrollment profiles once formal immunization coverage, birth attended by skilled staff, as school starts. Graph 15, panels A and B, sketch such well as outcome indicators such as the under-5 mortality enrollment profiles both between the non-poor HINOG rate all show that Turkey, given its level of income, is and poor LINOG children as well as between poor boys somewhat underperforming when compared internationally. and poor girls in the LINOG only in Turkey as of 2004. Graph 16a visualizes such relationship by plotting the 38 Sharp contrasts emerge, both by opportunity group and under-5 mortality rate against constant GDP per capita. poverty, as well as, almost more pronounced, by gender. The curve indicates a ‘predicted’ outcome – and we see Such gender differences in education access, beyond that Turkey finds itself with a higher than predicted under- and above the opportunity setting of households, once 5 mortality rate. Similar results can be observed when we again emerges as one of the core development challenges compare immunization coverage or attended births (the 37 for Turkey. latter shown in Graph 16b). Graph 16a: Under-5 Mortality Rate (latest available year, per 1,000 Graph 16b: Births Attended by Skilled Staff (latest available year, live births) percent) Source: World Development Indicators (2009) Source: World Development Indicators (2009) 37 See, e.g., ERI (2009). 38 Graph 15 does not include observations for Sub-Saharan African countries. Turkey: Expanding Opportunities for the Next Generation 16 5. Expanding Opportunities for the Next Generation: Early Childhood Development Policies This Section of the Report looks at pro-equity policy – 55. This Section focused on one such pro-equity policy, and focuses on one policy which has proven, in early childhood development (ECD). Given the strong international evaluations, to be highly cost effective and influence of the intergenerational transmission of inequity of particular benefit to the most disadvantaged groups: in Turkey observed above, supporting disadvantaged early childhood development. Given the observed strong children today could result in one of the most effective linkages between inter-generational opportunities and investments into Turkey’s future. ECD policies not poverty, on the one hand, and children’s health, cognitive only attack poverty today – through an indirect impact development and educational attainment outcomes, on on female labor force participation as well as reduced the other, early childhood development policies have an expenditures of childcare and preschool services – but important potential in Turkey to weaken the link between they are also the key for tackling the intergenerational birth circumstances and outcomes. Turkey is transmission of inequities and poverty. Given the large implementing innovative reforms that focus on children, positive externalities and high social as well as economic including the future pivotal role of family doctors. We returns from investing in the early years, many countries find, though, that funding priorities of social policies around the world have made ECD investment a top are concentrated on other age groups in Turkey today priority in their longer-term growth and development and that coverage of effective ECD programs, especially agenda. of the most disadvantaged poor children, is low. We find potential economic and social benefits, both at the ECD – Why and How? individual and societal level, to be very substantial. 56. The importance of Early Childhood Development 54. Public policies that would reduce inequalities of is due to brain development in the early months and opportunity in society are broadly those that would years of childhood affecting physical health, learning 40 attempt to weaken the link between circumstances, which and social behavior throughout life. In the early years, people cannot be held accountable for, and outcomes. a child develops all the basic brain and physiological A wide array of such policies exist, starting from building structures upon which later growth and learning depend. human capital through providing skill building and life- For instance, stunting in early childhood has been proven long learning opportunities for disadvantaged groups to to significantly impact cognitive development; iodine supporting productive asset creation by, e.g., addressing deficiency can lead to mental retardation; and insufficient capital market imperfections or income transfers which cognitive stimulation reduces learning abilities. Ensuring would offset original disadvantage. These could also adequate nutrition, health and cognitive stimulation in be policies that open opportunities by connecting people the very early months and years, raises returns to later 41 to markets, using public infrastructure investment to child investments significantly. The older a child gets overcome geographical poverty traps, and facilitating with developmental delays, the more intense, costly access for people to move and benefit from higher and less likely it becomes to recover the child back on 39 opportunity areas. his/her normal developmental trajectory. 39 World Development Report (2009). 40 See Heckman (2008), Engle et al (2007), Grantham-McGregor et al. (2007), Walker et al (2007), and Young and Richardson (2007). 41 Engle et al (2007). Turkey: Expanding Opportunities for the Next Generation 17 57. The international evidence on the effectiveness of such programs early on in their life have received ECD policies is growing – and a number of these derive higher scores on intellectual aptitude tests than non- 45 from the successful piloting of core ECD interventions participating children. and their evaluations in Turkey itself. For example, a long term study of the impact of the Mother Child 59. Internationally, several common characteristics of Education Program (MOCEP) in Turkey followed ECD interventions are linked to their success in reducing children who had participated in the first rounds of the disadvantage and breaking the intergenerational 42 46 program in 1982 until adulthood, 22 years later. It transmission of poverty and inequity. First, highest finds that children who benefited from the program returns, both social and economic, stem from focusing were, compared to a control group, more likely to have and reaching the most disadvantaged children. Second, graduated from high school and even university, and reaching disadvantaged children as early as possible, were more likely to be employed when compared to significantly before they reach kindergarten or pre-school peers. Such results are in line with international evidence readiness, is of highest impact and most cost-effective. from emerging market economies: In Colombia, Third, reducing iodine and iron deficiencies, supporting Argentina and Brazil, studies have found that children cognitive stimulation, and preventing stunting through who participate in ECD programs repeat fewer grades health and nutrition interventions require a holistic and progress better through school than non-participant approach to ECD that ideally encompasses health, children with similar backgrounds. 43 nutrition, cognitive stimulation and parental involvement 47 interventions together. Lastly, most cost effective 58. More recent studies emphasize that effective early interventions provide direct learning experiences to childhood programs not only have an impact on their disadvantaged young children and their families, are of school readiness and social and emotional development longer duration, high quality, and high intensity. For 44 but may also impact the level of children’s intelligence. advanced OECD countries, some of these policy lessons Evidence from programs in Cali – Colombia, Jamaica’s – as well as input indicators – have recently been First Home-Visiting Program, and Peru’s PRONOEI summarized by UNICEF in an international benchmarking program shows that children having benefited from assessment, provided in Box 1. Box 1: International Benchmarking League Table for Early Childhood Development UNICEF (2008) recently published an international benchmarking league table for economically advanced countries whereby quality and access standards for ECD programs are compared across countries. Such international targets include both structural policy measures (e.g., existence of national policy to support disadvantaged children) as well as coverage indicators of ECD programs (e.g., subsidized and regulated childcare being available for 25 percent of children before the age of three; subsidized and accredited early education services reaching 80 percent of four year olds) as well as input indicators for ECD programs (e.g., minimum staff to children ratio of 1:15 in preschool education; percentage of staff in accredited early education services educated with relevant information; 1 percent of GDP spent on early childhood services, excluding health services). Most of these benchmarks are relevant for OECD countries with advanced ECD programs and policies – e.g., over two thirds of young children in the 3 and 4 age group are now enrolled in preschools and kindergartens across OECD countries as a whole so setting a target for 80 percent becomes achievable with appropriate policies and planning. Benchmarking Turkey against such indicators would not be appropriate at this point but they could serve as long-run goals for the country given its renewed emphasis on reaching children early through its social policies as documented in major Government plans and strategies. 42 Kaðýtçýbaþý et al. (2005). 43 World Bank (2006), Myers (1995), Young (2002). 44 Heckman and Cunha (2007) 45 World Bank (2006), Chapter 7. 46 Engle et al (2007). 47 Evans, Myers, and Ilfeld, 2000. Turkey: Expanding Opportunities for the Next Generation 18 ECD Policies and Programs in Turkey cognitive development of children and increase school 49 preparedness. While parent and care-giver training 60. Today, Turkey is implementing a number of early programs provide a very important starting point for childhood development programs through public as expanding ECD through home-based services in Turkey, well as non-governmental delivery channels – with their coverage still remains modest. We estimate the specific targets having been defined for further roll-out share of children below age six reached through all during the coming years. We can distinguish four broad family training programs in Turkey (MOCEP, UNICEF’s types of programs: My Family Program and TAPF) to be around 3.5 percent (see Technical Endnote 7). 61. Center-based Pre-Schooling. Formal pre-school (center-based) attendance has increased significantly 63. Growth and psycho-social monitoring. The country, over the past five years, according to official figures through the Ministry of Health, is rolling out an now reaching about 50 percent of all children at age 6, innovative, integrated mother-child health monitoring and about 31 percent of children in the 5-6 year age and service, centered on the expanded responsibility of group. Public awareness drives, such as the “7 is Too family doctors. The model includes a multi-disciplinary Late ” campaign supported by the Mother Child assessment of child development and child needs in Education Foundation (AÇEV), have contributed to early ages which captures nutritional, health as well as mobilizing parents and public resources around early cognitive and psycho-social developments. The tool, childhood education. An infrastructure push fueled a digitalized scorecard consistent with World Health expansion of preschool classroom capacity within Organization guidelines, aims to track children starting existing primary schools. While supported at the national with the pregnancy of the mother. For each child, level, funding mostly depends on mobilizing local information on immunizations, growth performance, resources at the provincial level – hence local demand as well as cognitive and psycho-social developments, and prioritization become the main drivers of pre-school is regularly recorded. For this, the family doctor collects coverage. Teacher salaries and infrastructure costs are information from both mother and child on a regular covered directly by the Ministry of Education but private basis. Within the Ministry, the Directorates of Mother user fees also contribute to overall financing, covering, and Child Health and Primary Health closely coordinate for example, nutritional supplements. Such private fees policy and program development on the basis of the can vary between 50 TL and 200 TL per month and per integrated monitoring tool. Program implementation child – a potentially substantial cost for poor families started originally in Bursa in 1996 and is being currently considering that the poverty line for a family of four is implemented as pilots in 54 provinces, covering 108,000 around 700 TL today. pregnant women and 686,650 children in the 0-6 age group (equivalent to 8.3 percent of children in Turkey 62. Parenting programs. Turkey is recognized as an in the 0-6 age group). It is to be rolled out universally international pioneer in the provision of home-based with the family doctor program until the end of 2010. early childhood education and preschool programs. In the program, the family doctors are responsible for These are implemented by the Ministry of Education the immunization and growth monitoring of all children through the Non-Formal Directorate in Adult Education in their jurisdiction and receive per-capita incentive Centers (with financial support from the European payments on the basis of the number of families mapped Commission) and they are delivered in cooperation to their practice. 50 with the Mother Child Education Foundation and UNICEF. The programs support mothers (and now 64. SHÇEK Community Centers. Lastly, Turkey has 48 also fathers) as caregivers and are geared to improve developed a specific outreach to poor and disadvantaged 48 See AÇEV (2007), Kaytaz (2005) and Kaðýtçýbaþý (2005). The Government aims to reach 30 percent of parents and caregivers through the expansion of parenting programs in the coming years. 49 The Family Training programs by UNICEF (The My Family Program) trains mothers of children below age 6. The training programs are complemented through center based preschool training for children. The Father Training Programs implemented by AÇEV, for fathers of children ages 2-10 encourage fathers to engage more closely with the development of their children. More recently, the Directorate of Non-Formal Education and Apprenticeship in MONE has started to implement “care giving” courses for young women who would like to take up jobs at homes, day-care centers or preschools as caregivers. 50 Staff elaboration based on Ministry of Health Information. Turkey: Expanding Opportunities for the Next Generation 19 children through its Social Services and Child Protection The current regulatory framework for daycare centers Agency (SHÇEK). SHÇEK operates both, what it calls, and pre-schools emphasizes the quality of infrastructure ‘protective’ services for orphans and children in need as (i.e., the availability of an open air space, sleeping room well as ‘preventive’ services through Community Centers with specific area requirements, kitchen, medicine room and family counseling programs. Community centers etc) as well as training requirements for staff working are multi-purpose facilities located in disadvantaged in the centers. Reviewing such current infrastructure communities – often they deliver a range of services, and staff training requirements with the aim of facilitating the accreditation and expansion of services, coupled from family training programs, vocational training with offering public financing for those private providers workshops, counseling for families to providing child- 51 that serve in disadvantaged areas, could have the potential care options for course attendants. The number of to expand day-care opportunities for poor children and community centers across the country is 81 with an their mothers. estimated reach of around 40.000 children (see Technical Endnote 7). Governance and Coordination Private and Community Provision 66. The governance of ECD programs rests with various institutions in Turkey. The Ministry of National 65. Scaling up the provision of ECD through low-cost Education assumes a national coordination role. The community models as well as private providers could ‘Inter-sectoral Child Council’ is entrusted with the be an interesting tool for expanding ECD services to coordination of the ECD policy within the framework the most disadvantaged children in Turkey. As of 2007, of the UNICEF-Government of Turkey program. Given less than 1 percent of the 5-6 year old group benefited the various actors delivering ECD services across the from private centers and only about 6 percent of total public, voluntary and private sectors, such coordination enrollments of children between 4 and 6 years of age is essential. The Council is attended by various were at private schools. Hence, community-driven and ministries. While the Ministry of Education assumes private day-care and pre-school providers are rare a coordinating role, the quality control of service delivery 52 albeitinnovative and promising models exist (see Box 2). is the responsibility of the respective line agencies. Box 2: Community-Based ECD Provision in Turkey One community-driven model of private provision has been developed by KEDV (Kadýn Emeðini Deðerlendirme Vakfý), an NGO empowering women through networks and cooperatives. KEDV facilitates the establishment of cooperative day-care centers (under the name of “Women and Children Centers”) in poorer urban neighborhoods of Turkey. Its model provides a low-cost solution for affordable day-care provision in poorer neighborhoods whereby women in the communities come together and self-finance day-care services for their children. KEDV staff train and support “neighborhood mothers” who volunteer at the centers; the cooperative also has the possibility to hire professional preschool teachers to work in the day care centers. Members of the cooperative contribute to costs on a sliding scale according to their ability to pay, such that over time, the neighborhood nurseries become entirely self-financed and self- 53 run by the communities. Such community driven initiatives may provide a method of expansion of ECD services in low income urban neighborhoods of Turkey where otherwise private and public child care options are limited or too costly. In many countries, particularly in Latin America, such community-driven informal networks have become pillars of expansion of ECD services, often supported through public financing within a system of accreditation and close supervision (Young 2007). 51 SHÇEK currently operates 81 Community Centers across the country, 26 Women Shelters and 42 Family Counseling Centers. 52 The supervision of growth monitoring and ECD health services delivery rests with the Ministry of Health which also oversees family doctors and primary health care facilities. Similarly, the Social Services Agency SHÇEK accredits private providers of day-care and crèche facilities (albeit this responsibility will shift to the Ministry of Education). So far, the Ministry of Education has regulated the educational content of private nurseries and preschools, while SHÇEK has provided license supervision. In the future, SHÇEK’s license supervision authority will be passed to MONE. 53 Social Policy Forum (2009). Turkey: Expanding Opportunities for the Next Generation 20 67. Institutional reforms that allow for effective planning in place that have served as examples to the world, we and coordination of activities across line ministries and estimate that coverage as well as reach among various government levels are often at the heart of successful poorer,disadvantaged children is low for most ECD ECD campaigns. Recently, Chile has undertaken a programs in Turkey today. Graph 17 shows approximate comprehensive institutional reform to implement its overall coverage rates, distinguished by child develop- program Chile Crece Contigo (Chile Grows With You). ment phase and intervention type. Apart from birth The program is, de facto, a universal and integrative child attendance and immunization (90 percent) and formal, insurance system that covers all children in Chile from center-based pre-school education (around 30 percent conception onwards with services that ensure their healthy for 5-6 year old group) no program currently reaches growth and development. Coordination and financing more than 10 percent of children in the relevant age functions, through the Chilean planning ministry, are group. As we had seen in Section Four, even the high separated from implementation which remains with coverage programs such as immunizations or birth responsible line agencies and specialized institutions. attendance tend to reach primarily non-poor children from better opportunity groups. For the other programs, Coverage and Reach among Disadvantaged Children even if these were focused on reaching the poor and disadvantaged children, under-coverage of children at 68. While Turkey has many innovative ECD programs risk would still be significant. Graph 17: Coverage of Various ECD programs in Turkey by Age Category Source: Staff elaboration. See Technical Endnote 7 for details 69. Over the past years, Turkey has placed much emphasis of capacity was funded largely at the provincial level. on expanding pre-school access for six year olds. Overall Further targets have been set: Turkey now aims for the enrollment shows a strong and sustained increase over pre-school enrollment rate to reach 100 percent in 30 provinces by 2011. As Graph 18 shows for 2008, further the past years, rising from less than 15 percent in 2003 strong increases in preschool enrolment rates would be to more than 30 percent in 2008 of the 5-6 year old group. necessary to approach levels observed in comparator This was possible through formation of strong partnerships countries – or approaching targets specified by the between central and local levels as the expansion OECD. Turkey: Expanding Opportunities for the Next Generation 21 Graph 18: Pre-primary Gross Enrollment for 36-72 months, percent 71. Matching the household-level results, the provincial (2008) distribution of access rates to pre-school is also higher in better-off provinces today. Graph 20 compares provincial pre-school access rates with the level of human development (as measured by the Human Development Index, compiled by UNDP) at the provincial level. In line with the Koç data – a mild negative correlation can be witnessed, signaling that pre-school enrolment rates tend to be higher in better- off provinces. Graph 19: Preschools and Daycare Access, by socioeconomic status and rural/urban location (2008) Source: World Development Indicators (2009), MONE 2008-2009 enrollment data 70. The combined public and private supply of pre- schools, kindergartens and childcare options – which are dominated by publicly provided pre-schools – is, today, strongly tilted to better-off areas. Household level data on such availability, collected by Koç University (Graph 19), show that less than 20 percent of mothers from low socio-economic status have access to such facilities in their neighborhood while more than 50 percent of mothers with a high socio-economic status have such local access. The same disparity reappears along the rural- urban divide. Local elementary school Note: Socio-economic status is defined as an index that combines income, household assets and the educational level of adults in the household. availability, on the other hand, is more evenly distributed Source: The Study of Early Childhood Developmental Ecologies in Turkey (2008), Koç across these dimensions. University. Graph 20: Pre-School Enrolment and Human Development Index, by Province Source: Staff Elaboration based on MONE (2008-2009 academic year) and UNDP (2004) data. Turkey: Expanding Opportunities for the Next Generation 22 Public Funding sectors. Highest is the funding of health programs, mostly for immunization and growth monitoring and 72. While funding for Early Childhood Development services through family doctors and primary care policies and programs has gradually increased in Turkey, institutions. Education expenditures for the age group a relatively small share of central public social are low since only about four percent of the total central expenditures is directed towards children today. Our public education budget accrues to publicly funded pre- calculations, approximations only, are based on a detailed 57 school and family/parent training programs. Lastly, classification of all central social expenditures in central social protection expenditures are also relatively education, health, social protection, excluding all health, modest for the age group and largely consist of funding unemployment and pension contributions to the social 54 of the child-focused conditional cash transfer payments security institute. Given both program nature as well – which currently reach more than 3 million children as available estimates as to the group-specific nature with modest transfers – as well as funding for the most of funding), we derive estimates for age specific social disadvantaged children provided through the Social expenditures. The expenditures included here only Services and Child Protection Agency (SHÇEK). reflect central government spending and do not include local government expenditures or private expenditures at the household level. As Graph 20 shows, of the total 75. Given the young demographic structure of the central, social expenditures, about 6.5 percent benefited Turkish population today, per capita funding levels the youngest age group in society while the largest share between age groups diverge stronger. Graph 21 shows of funding accruing to the age group between 45 and per capita funding across the entire population – hence 65, primarily due to the Treasury-financed (non- spreading funding across those that benefit from specific contributory) portion of pension and health expenditures. programs and those that do not. As shown, estimated Overall, Turkey spends about 0.5 percent of its GDP central, non-contributory funding levels for each child on the 0-6 age group – considerably less than established in the 0 to 6 age group in 2008 reached approximately international Early Childhood Development benchmarks 55 Graph 21: Distribution of Social Expenditures in Turkey for OECD countries. by Age Group (2008) 73. This finding changes little if we make different assumptions as to who benefits from social transfers within households. Rather than, as above, assuming that income transfers (e.g., all income-based social protection transfers) benefit the recipient only, we estimated the age-specific expenditure incidence in the alternative extreme, distributing all transfers equally to all household members. In such case, children in the 0-6 year old group receive 2.8 percent of pension benefits 56 in Turkey. With this revised calculation, the distribution of public expenditures by age group changes only slightly and the overall share of funds reaching the youngest age group would still be only about 7 percent. Source: Ministry of Finance Central Budget Data, staff calculations (see Technical 74. Early childhood funding varies significantly by Endnote 8 for details). 54 The budget figures exclude all health, unemployment and pension contributions by households to the social security institute. See Technical Endnote 8 for details. 55 UNICEF (2008) establishes a benchmark of one percent of GDP for ECD programs, excluding health. The comparable amount for Turkey without health expenditure amounts to roughly 0.1 percent of GDP. 56 Children would receive 2.8 percent of pension benefits, 12.2 percent of social assistance benefits through the SYDGM and 6.6 percent of non- contributory old age benefits. However, since the overall size of these benefits is modest, making these different assumptions about the intra-household distribution does not alter the original age-incidence significantly (see Technical Endnotes 8 and 9). 57 As an approximation, we assign 1/3 of the total budget of the Non-Formal and Apprenticeship Education Directorate to parent and family training activities. The salaries of full-time preschool teachers who teach within primary schools are also included in the “preschool” budget in this calculation, although they are paid from the budget of the DG for Basic Education. Turkey: Expanding Opportunities for the Next Generation 23 650 TL, mostly related to immunization, post-natal care coordination could be organized. Also, significant work and growth monitoring programs. A slight drop for the has been carried by partner organizations to assess the 6 year olds is followed by a steep increase for those potential costs of expanding parenting and pre-school children in primary and an even steeper increase for education. Hence, this Report focuses on the benefit those in secondary school – as well as considerable side. As evidenced by international studies, long-term funding provided also for students in university which, benefits from high-quality early interventions programs on a funding per beneficiary basis, is considerably higher for disadvantaged children include higher verbal and than for other education levels. Social expenditures for mathematics achievement, greater success at school (less the working age population drops to around 300 TL grade repetition, higher graduation rates), higher before a steep rise occurs at the current minimum employment and earnings, better health outcomes, less retirement age in the mid-forties. High health and pension welfare dependency, and lower crime rates. 59 expenditures through middle- and old age lead to a significant tilt of overall social expenditures to the right 77. An impressive body of evaluations illustrates such of our age scale – and resulting, on a per capita basis, in benefits for several Turkish programs. Data from a ratio of funding between the ‘over-45 year’ age group controlled experiments reveal that children who take 58 and those between 0-6 year olds of almost four to one. part in pre-school programs achieve roughly one more year of formal schooling (excluding the pre-school 60 Graph 22: Per Capita Social Expenditures by Age Group in Turkey attendance year) then children who do not participate. (Central Government, Excluding Social Security Contributions, 2008), With more schooling comes higher incomes, a higher ‘000 TL likelihood of working in the formal sector and, linked, higher contributions to public revenues. The returns to investment in one year home-based pre-school education through the MOCEP program only, have been calculated 61 at a minimum to be 1:2 and reaching up to 1:6. Even when looking at family training programs in isolation returns of between 20 percent and 100 percent for every TL invested result. Apart from high private benefits – and preventing the participating children from falling into poverty later in their life – UNICEF estimates that public investments in such family training programs pay for themselves through higher formalization and 62 tax collection within roughly twenty years. Note: Staff calculations. (See Endnote 8 for explanations.) 78. Improving the opportunities of Turkey’s children Early Childhood Development Benefits today would have significant economic and social benefits. This Report uses two simulation models to 63 76. This Report focuses on assessing the economy-wide assess potential benefits of selected ECD policies. The benefits of ECD policies. Without doubt, costs of expanding first one focuses on the impacts of educational attainment 64 early childhood development services are substantial. We stemming from pre-school/parenting in Turkey. We did not undertake a full and comprehensive costing exercise use a simulation model that examines how poverty, for this Report as this would very much depend on the incomes and female labor force participation would be model of Early Childhood Development programs being different today for the generation of 20-39 year offered – including how, where and who would deliver olds if all of them had assisted pre-school/parenting the various dimensions and how effective and efficient programs when they were six years old. Hence, given 58 See Technical Endnote 8. 59 Engle et al (2007), p. 231. 60 Ertem et al (2006), Kaðýtçýbaþý et al. (2001), Kaytaz (2005), Ozmert et al (2005) and Yýlmaz (2007). 61 Kaytaz (2005). 62 Yýlmaz (2007). 63 Both models concentrate on assessing direct impacts only. Hence, they do not assess overall general effects which would have to include many more behavioral functions. 64 Sunar, Bekman and Kaðýtçýbaþý (2001) and Kaytaz (2005). Turkey: Expanding Opportunities for the Next Generation 24 that evaluations in Turkey have shown a net increase work or choose their self-employment with some sectors of schooling by about one year for these children, we more prone to absorb higher-skilled employees than hypothetically raise the educational attainment of this others. Second, increases in education tend to be age group by this one year. We then consider impacts associated with a reduction in the fertility rate of women, of influence which such higher educational attainment: leading to changes in the availability of household per first, education is linked to both participation in the capita income and consumption. Lastly, the higher labor market and occupational choice. Using observed educational attainment itself has a direct impact on relationships, participation rates, especially for women, earnings and consumption given that one more year of increase with higher educational attainment. In parallel, schooling is associated with a return that we can calculate education levels also influence in which sectors people from historical relationships. Box 3: Female Labor Force Participation and Affordable Child Care Options in Turkey In addition to the direct impact of ECD interventions on children, interventions that have an affordable child care component can have beneficial impact on mothers’ participation in the workforce. Today, Turkey has a low and declining level of female labor force participation: with 25.5 percent as of 2009, the country has the lowest female labor force participation among OECD and EU-19 countries. Turkey’s 9th Development Plan targets an increase of female labor force participation in Turkey to 29 percent by 2013. To date, the Government has enacted measures to increase the demand of firms for female employment – such as partial subsidies offered to enterprises that hire additional, first-time formal worker women into their companies (Employment Package passed in May 2008). There is a new and expanding body of evaluations that look at female labor force participation rates in the presence of an expansion of affordable ECD services. A study in Argentina estimates the effect of a large-scale increase in the availability of free public pre-school on maternal employment to be between 7 and 14 percent (Berlinski and Galiani 2007). Using a regression discontinuity design, another study in Argentina estimates a 13 percent difference in labor market participation between mothers whose youngest child just made the age cutoff for pre-school eligibility and those whose youngest child just missed that age cutoff (Berlinski et al, 2008). A companion Report to this study (World Bank, 2009a) finds that in Turkey, reducing the opportunity cost of working for women in urban areas, e.g., through providing increased availability of day-care services, could also have potentially large and immediate positive female labor force participation impacts. In Turkey, as in many other countries, marriage and child birth are negatively correlated with women’s labor supply decisions. Of the women in the 20-65 year age group, who have ever been married, more than half state being a housewife as a reason for not working while 9 percent state “taking care of children” as the main reason for not working (Turkey Demographic and Health Survey, 2004). Child birth is a significant contributor in Turkey to lower participation in the labor force, particularly for women living in urban areas, where informal networks of child care are weaker. A multivariate analysis finds that the birth of the first child is correlated with a reduction in the probability of labor force participation by about 12 percentage points for high skilled women and a reduction of around 10 percentage points for low skilled women in urban areas (Aran et al., 2009). Given that access to affordable child care services are very limited in Turkey, women – especially in urban areas after child-birth – have limited options for continuing to work. Because of the high costs of child-care services in the country, the opportunity cost of working and the reservation wage of women increases after child birth. While some women choose to stay home and care for their children, the absence of affordable child care options inhibits labor force participation for women that would like to continue working. This can especially be the case for young migrant families in urban areas who can no longer rely on an extended family network for child-care (Social Policy Forum, 2009). Increasing options of good quality day-care can, in addition to providing positive cognitive and development stimulus for young children, alleviate a supply constraint for women’s labor force participation in Turkey. Turkey: Expanding Opportunities for the Next Generation 25 6. Reflections 82. This Report examined equity through the lenses of into the early child development outcomes of their opportunities. We found that inequalities of opportunity grandchildren, today’s young generation in Turkey. Child are substantial and that what is termed circumstances development trajectories diverge early on depending on here, i.e., exogenous factors that no individual can opportunity background as well as poverty status. One, influence, are an important determinant of life of several, pro-equity policies focuses on reaching the trajectories. We also found that the intergenerational most disadvantaged children early on in their life, ideally transmission of such inequities is pronounced, with the before birth. Today, such early childhood development socio-economic status of grandparents mapping closely policies in Turkey reach relatively few of the children 65 Using the same methodology to estimate the impact of a larger increase of female labor force participation, i.e. to 40 percent, we find an aggregate income effect of almost 12 percent and a poverty reduction impact of 17 percent. Turkey: Expanding Opportunities for the Next Generation 26 most in need, while public resources are largely benefitting of the child’s life is particularly important given that other age groups. women’s educational attainment remains low, adult female literacy is not universal and mothers often remain 83. We end the Report with a number of reflections for the primary caregivers of children. the public social policy debate. First, to improve equity in society, opportunities for disadvantaged children would 85. Third, international evaluations show that the most need to be expanded which will necessitate reviewing effective way to reduce the influence of circumstances the current functioning and financing of the Turkish on opportunities is to provide effective support to the welfare state. Given the equity aspirations of the Turkish most disadvantaged children first. The concept of equality people which we reviewed in this Report, the society at of opportunities employed in this Report goes beyond large appears open to enter such reflection. A host of creating equal access – it implies that the most pro-equity policies exist and one of them, early childhood disadvantaged are reached first and more intensively than development policies, have proven significant in many less disadvantaged children so that they can improve countries in weakening the link between circumstances their life chances. If this concept resonates in Turkey, and individual life chances. Current public, non- a strategy for roll-out would then need to set targets contributory social expenditures reach children in their detailing how, and how many, of the most disadvantaged early years to a limited extent. To create fiscal space children can be reached. Currently, vital child services that would allow programs for disadvantaged children – from nutritional support, health attention as well as to expand, the financing and societal transfers to old age cognitive stimulation – largely benefit less disadvantaged insurance would need to be re-examined. children. While a discussion on how ‘disadvantage’ can be defined will be necessary, this Report suggests that 84. Second, public policies for the most disadvantaged two factors alone, child poverty and parental education, children would need to complement the existing informal solidarity networks in the country. These informal safety are core determinants of opportunities. nets, as strong as they might be, do not seem to have been able to offset the disadvantages of children born 86. Lastly, civil society, community and private initiatives into specific circumstances, especially for girls. Turkey will need to complement public efforts in expanding the has a traditional and strong communal and family supply of services for the most disadvantaged children. solidarity which is often described as being one of the Turkey’s innovative and inspirational experiences of 66 main pillars of societal functioning. But given the ECD service delivery through non-governmental channels strong inter-generational transmission of inequity observed is looked at across the globe with much admiration. In here, such communal and traditional ties would at least parallel, overall coverage of such delivery is very low. need to be complemented by an integrative and inclusive A social consensus to support the most disadvantaged policy for the most disadvantaged children. Families children would require a compact between private, public, could then be supported to improve their children’s civic and community actors to create an appropriate healthy early development and learning. Supporting enabling environment for the availability of high quality families through ECD programs in the early years and integrative support for disadvantaged children. 66 See, for example, Ayata (2004). Turkey: Expanding Opportunities for the Next Generation 27 References AÇEV (2007), “Seven is Too Late Campaign Evaluation Cunha, F. and J. Heckman (2007), The Technology of Results”; AÇEV (Mother Child Education Skill Formation, NBER Working Papers 12840, Foundation), Istanbul. National Bureau of Economic Research. Aran, M., Uraz, A., Husamoglu, M., Sanalmis, D., and Education Reform Initiative (2009), “Inequality of S. Capar, (2009). “Recent Trends in Female Labor Opportunities in Education”, Istanbul. Force Participation in Turkey”; SPO-World Bank Welfare and Social Policy Working Paper No. 2, Engle, P., Black, M., Behrman, J., Cabral de Mello, M., Ankara. Gertler, P., Kapiriri, L., Martorell, R., Young, M.E. and the International Child Development Steering Aran, M., Demir, S., Yazýcý, H., and Sarýca, O. (2009), Group (2007), “Strategies to Avoid the loss of “Poverty and Inequality Changes in Turkey”; SPO- developmental potential in more than 200 million World Bank Welfare and Social Policy Working children in the developing world”; LANCET 369: Paper No. 1, Ankara. 229-242. Ayata , S.(2004), “Poverty and Policy in Turkey”; Ertem, I., Atay, G., Bingoler, E., Dogan, D., Bayhan, Middle-Eastern Technical University, Ankara. A., and D. Sarica (2006), “Promoting child development at sick child visits: A controlled trial Baydar, N. (2008), “Study of Early Childhood to test the effect of the home environment of young Developmental Ecologies in Turkey”, Koç children”; Pediatrics 118: 124-131. University, Istanbul. European Bank for Reconstruction and Development Bekman, S. (1998), “A Fair Chance: An Evaluation of and World Bank (2007), “Life in Transition Survey”, the Mother-Child Education Program”, Istanbul. London. Berlinski, S., and S. Galiani (2007), “The effect of a Evans, J., Myers, R., and E. Ilfeld (2000), “Early large expansion of pre-primary school facilities on Childhood Counts”, The World Bank, Washington preschool attendance and maternal employment”; DC. Labour Economics, 14: 665-680. Ferreira, F. and J. Gignoux (2009), “Inequality of Berlinski, S., Galiani, S., and P.J. McEwan (2008), Opportunity for Education: The Case of Turkey”, “Preschool and maternal labor market outcomes: SPO-World Bank Welfare and Social Policy Evidence from a regression discontinuity design”, Working Paper No. 4, Ankara. IFS Working Papers W09/05, The Institute for Fiscal Studies, London. Ferreira, F., Gignoux, J., and M. Aran (2009), “Inequality of Economic Opportunity in Turkey”, SPO-World Bourguignon, F., Ferreira, F., and Ph.G. Leite (2008): Bank Welfare and Social Policy Working Paper No. “Beyond Oaxaca-Blinder: Accounting for 3, Ankara. differences in household income distributions”, Journal of Economic Inequality 6: 117-148. Filmer, D. and L. Pritchett (2001), “Estimating Wealth Effects without Expenditure Data – or Tears: An Bourguignon, F., Ferreira, F. and N.Lustig (eds.) (2005) application to education enrollment in stats of India”, “The Microeconomics of Income Distribution Dyna- Demography 38 (1): 115-132. mics in East Asia and Latin America”, The World Bank and Oxford University Press, Washington Government of Turkey (2007), Ninth Development DC. Plan, State Planning Organization, Ankara. Turkey: Expanding Opportunities for the Next Generation 28 Grantham-McGregor, S., Cheung, Y., Cueto, S., Paes de Barros, R., Ferreira, F., Molinas Vega, J., and Glewwee, P., Richter, L., Strupp, B. and the J. Saavedra Chanduvi (2009), “Measuring Inequality International Child Development Steering Group of Opportunities in Latin America and the (2007), “Development potential in the first 5 years Caribbean”, Latin American Development Forum for children in developing countries”, LANCET Series, Washington DC. 369: 60-70. Roemer, J. E. (1998), Equality of Opportunity, Cambridge, MA, Harvard University Press. Heckman, J. J. (2008), “The case for investing in disadvantaged young children”, in: First Focus Seifer, R., and A. J. Sameroff (1987), “Multiple (Ed.), Big ideas for children: Investing in our nation's Determinants of Risk and Vulnerability”, In E. J. future: 49-58, Washington, DC. Anthony and A. Cohler (eds.), The Invulnerable Child, pp. 51-69, Guilford Press, New York. Heckman, J., and F. Cunha (2007), “The Technology of Skill Formation”, The Institute for the Study of Shonkoff, J., and S. J. Meisels (Eds.) (2000), Handbook Labor, (IZA), Bonn. of Early Childhood Intervention, Cambridge University Press, Cambridge. Hentschel, J. and M. Aran (2009), “A Methodology Note on Assessing Labor Market and Poverty Shonkoff, J.P., and D. A. Phillips (2000), From Neurons Implications of a Possible Growth Slowdown in to Neighborhoods: The Science of Early Childhood Turkey”, The World Bank, mimeo, Ankara. Development, National Academy of Sciences Presses, Washington, DC. Kagitcibasi, S., Bekman, S. and Z. Cemalcilar. (2005). Social Policy Forum (2009), Turkiye'de Cocuk Bakim “Continuing Effects of Early Intervention in Adult Hizmetlerinin Yayginlasmasina Yonelik Bir Oneri: Life”, AÇEV, Istanbul. Mahalle Kreþleri, Bogazici University, Istanbul. Kaytaz, M. (2005), “A Cost Benefit Analysis of Sunar, D., Bekman, S., and Ç. Kaðýtçýbaþý (2001), Preschool Education in Turkey”, ACEV, Istanbul. “Long-term Effects of Early Intervention: Turkish Low-Income Mothers and Children”, Applied Ministry of National Education (2009), Statistical Developmental Psychology: 333-361. Yearbook: 2008-2009, Ankara. Taymaz, E. (2009), “Growth, Employment, Skills and Myers, R. (1995), The Twelve Who Survive: Strengthening Female Labor Force”, SPO-World Bank Welfare Programs of Early Childhood Development in the and Social Policy Working Paper No. 6, Ankara. Third World, 2nd ed., High/Scope Press, Ypsilanti, Michigan. TUSIAD (2005), "Dogru Baslangic: Turkiye'de Okul Oncesi Egitim", Yayin No. TUSIAD-T/2005-05/396, Istanbul. O’Donnell, O., Doorslaer, E. v., Wagstaff, A., and M. Linde (2008), “Analyzing Health Equity Using UNICEF (2008), “The Child Care Transition: A League Household Survey Data”, World Bank, Washington table of Early Childhood Education and Care in DC. Economically Advanced Countries”, Innocenti Research Center, Florence. OECD (2008), Education at a Glance, Paris. Walker, S., Wachs, T. D., Meeks Gardner, J., Lozoff, Ozmert, E. N., Yukdakok, K. and S. Soysal (2005), B., Wassermann, G., Pollitt, E., Cater, J.A. and the “Relationship between physical, environmental and International Child Development Steering Group socio-demographic factors and school performance (2007), “Child development: Risk factors for adverse in primary schoolchildren”, Journal of Tropical outcomes in developing countries”, LANCET 369: Pediatrics 62: 25-32. 145-157. Turkey: Expanding Opportunities for the Next Generation 29 Wietzke, B. (2009), “Neighborhood Effects and Human World Bank (2009c), World Development Indicators Development: A Critical Review of the Literature”, 2009, Washington DC. London School of Economics, mimeo, London. Yýlmaz, H. (2007), “A Short Review of Parent Education World Bank. (1999), “Early Childhood Counts: Ways Programmes (0-6)”, UNICEF, Ankara. in which ECD Programs can address inequalities”, Washington DC. Young, M. E. (2002), “From Early Child Development to Human Development”, World Bank, Washington World Bank (2006), “World Development Report, DC. Development and Equity”, Oxford University Press, Washington DC. Young, M. E. and Richardon, L.M. (2007), “Early Child World Bank (2009a), “Female Labor Force Participation Development: from Measurement to Action”, World in Turkey: Trends, Determinants, and Policy Bank, Washington DC. Framework”, Report 48508-TR, Washington DC. Zaidi, S., Alam, A., Mitra, P., Sundaram, R. (2009), World Bank (2009b). “Investing in Young Children: “Satisfaction with Life and Service Delivery in the Why and How, Practical tools and evidence- Eastern Europe and the Former Soviet Union”, based recommendations", Washington DC. World Bank Working Paper 162, Washington DC. Turkey: Expanding Opportunities for the Next Generation 30 Technical Endnotes 1 The basis for the inequality of economic opportunity 2 The inequality of opportunity share can be computed analysis is the Demographic and Health Survey through a (i) parametric regression (which relates (data collected from December 2003 to May 2004) the wealth variable directly to the circumstance which recorded data for 10,836 households, variable and uses the degree of explained variance representative at the national level but also for five as the, implicit, opportunity share; or (ii) non- major regions (West, South, Central, North and parametric analysis that partitions the entire dataset East Regions). Information on basic socio-economic in different, identical circumstance groups and characteristics of the population was collected for calculates how much of the overall variance of all households’ members and all ever-married household wealth is due to variation between the women between 15 and 49 years old answered a conditional cell means. The two techniques show detailed questionnaire on demography and health; similar results for Turkey. Important to note is that in total, 8075 women provided information. This the inequality of opportunity share is a lower bound ‘ever-married women’s questionnaire’ included estimate of the inequality of opportunity share given information on the characteristics of these women that not all circumstance variables are captured in at birth – place of birth (rural/urban; region), her the household survey (e.g., parental occupation or father’s and mother’s education; maternal language, parental wealth). See Ferreira et al (2009) for a etc. Hence, the results presented in this Report are detailed explanation. related to the wealth distribution at the household level with the circumstance variables being derived 3 As explained in Ferreira et al. (2009), the parametric from the women in each household. For a more estimate provides a robust lower-bound estimate extensive discussion, see Ferreira et al. (2009). of the share of inequality of opportunity in total wealth disparity in Turkey. The decomposition of To assess the composition of the group of households this share into the contribution each individual that did not count an ever-married woman among circumstance variable makes is, however, dependent its midst, we analyzed the Household Budget Survey on the partial correlation of the circumstances with for 2004 as this survey also includes information omitted variables (i.e., the circumstances that were on employment and other variables of interest. The not included in the Turkish Demographic and Health households which did not have an ever-married Survey). See Ferreira et al. (2009) for a robustness woman between 15 and 49 years old represented test of the results with an alternative to the wealth 17 percent of the total population (very close to the indicator (imputed consumption at the household 20 percent obtained from the Demographic and level). Health Survey) but accounted for only 3 percent of all children below 15. A majority of the individuals 4 Using the exogenous circumstances, the entire living in households without an ever-married woman population can be divided into different groups by – 63 percent – represent households made up entirely their opportunity sets. E.g., one such group is made of members that are 50 or more years old and not up of all households whose ever married women working. Reverting back to the Turkey were born in the urban areas of the West region, to Demographic and Health Survey, we find that the parents who both had higher education diplomas, average size of households without an ever-married whose mother tongue is Turkish and who had more woman is 2.70, much lower than the mean size of than nine siblings. Ferreira et al (2009) distinguish the households in the TDHS sample (4.63 in 2004). 768 such different types. For each type, they Poverty is also less prevalent for the population calculate the mean wealth which they use to rank living in those households with a rate of 19.8 percent, all types. The ten percent of the population with against 29.8 percent for the population in the TDHS lowest wealth (conditional by type) are then termed sample households. the least advantaged decile; the ten percent with Turkey: Expanding Opportunities for the Next Generation 31 the highest wealth (conditional by type) the most force; (ii) for the informal sector outside of advantaged decile. Examining the composition of agriculture, we assume the average historic output these two groups produces an opportunity profile elasticity of 0.37, leading to a contraction of the which is the basis for Graph 7 in the main text. sector affecting 45,000 heads of households until 2010. Similarly, in line with historic elasticity 5 The adjustments and assumptions we have made calculations, agricultural informal employment in the simulations to assess the possible poverty would not contract Over the past ten years, impact of the economic slowdown are as follows: agricultural employment has decreased strongly, independent of output decreases or increases · Our starting point is the 2006 Household Budget (which leads to an actual calculation of an, Survey for Turkey. Each household has a (spatially adjusted) real per capita consumption insignificant, employment elasticity of 0.07). assigned with a (national) poverty line defined; However, in times of crisis, such long-term · To bring the 2006 data (the latest available survey) declining trend could well be halted or, potentially, to 2008 as the base, we apply the real sectoral even be reversed as already seen in the latest GDP per capita growth rates to the consumption aggregate employment figures of TUIK; (iii) we per capita variable (distinguishing between the select the actual heads of households losing their agricultural, industry and services/other sectors). employment in the two ‘contracting’ sectors The affiliation of each household to a sector is (low-wage formal and informal non-agricultural determined by the activity of the household head. sector) by random assignment; (iv) for all other For inactive or already unemployed household sectors (high-wage formal employment, informal heads, we apply the average GDP per capita agriculture, inactivity, unemployment), we growth rate; assume that the employment status of the · We now map households into 6 different households head does not change. categories, based on the employment status of · To arrive at impacts on household welfare, we the household head: low-wage formal sector use the partial correlation coefficient (holding employment (those household heads with social all other potential influencing variables constant) security coverage and earning up to 1.5 times the minimum wage); high-wage formal sector between consumption per capita and the (above 1.5 the minimum wage); informal sector household head being unemployed. The cross- working in agriculture; informal non-agricultural section regression using the 2006 HBS data sector, unemployed; and inactive. returns that, on average, an unemployed head of · We assess the potential impact of a growth household is associated with an 18 percent lower slowdown of minus five percent real GDP growth consumption per capita level of the household. in 2009 and minus one percent real GDP growth We apply this reduction to the population living in 2010; in those households whose household head lost · The employment impact of the scenarios is their employment in our simulation. determined by calculated average employment elasticities of output growth (Taymaz 2009). We For the other sectors, we make the following assume the following: (i) the formal sector assumptions: (i) for those households linked to the employment would contract with the historically high-wage formal sector, we assume that they are calculated weighted elasticity of 0.47. In the relatively well able to protect their welfare level in growth scenario, this would lead to an estimated real terms (in part because a good part of the workers 170,000 heads of households losing their primary would be unionized and/or in the public sector). We employment in 2009. We assume that these would be concentrated among ‘low-wage formal assume a reduction of half of the GDP per capita sector workers’ – Taymaz (2009) finds that during growth rate to apply to them; (ii) for households the last crisis in Turkey, by far the largest labor whose head is unemployed or inactive, we apply the market adjustment occurred for these workers. full GDP per capita growth rate; (iv) the remaining Note that we assess only the labor market impact households (those remaining in the informal non- on household heads and not of the total labor agricultural sector after the above adjustments Turkey: Expanding Opportunities for the Next Generation 32 are made) would then experience a decline in their tracking data stem from the Ministry of Health consumption per capita level which ‘adjusts’ – the Primary Health Care DG (Mental Health Unit) and residual between the different assumptions for the refer to December 2008. Coverage of parent training other groups and overall GDP per capita growth. programs originates from the Ministry of Education Apprenticeship and Non-formal Education DG and Changes in consumption per household can then assumes that each parent reaches 2 children in the be analyzed by comparing the 2008 projected 0-6 age group, on average. Public and private consumption level of the household with the preschool coverage rates for 4-6 year old group simulated consumption per capita level in 2009 and stem from MONE. Preprimary education coverage 2010. We use this comparison to derive a profile data for 2008-2009 academic year refer to the sum of households that are at risk of falling into poverty of public nursery classes and public kindergartens. as a consequence of the economic slowdown. Primary school enrollment rates are from MONE 2008-2009 primary school enrollment data. SHÇEK 6 Definitions for variables (derived from Turkish community centers coverage is calculated as 81 Demographic and Health Survey) included in Table community centers each serving an estimated 4 are as follows: Antenatal care is defined as a number of 500 children in the age group (81 x 500 pregnant woman receiving care provided by a doctor = 40,500 children). or a nurse/midwife through at least one visit. This share is calculated for all most recent births for 8 The basis for the calculation of central public social ever-married women who had a live birth in the expenditures by age group is the functional breakdown five years preceding the survey. Birth certificate: of expenditures published by the Ministry of Finance the child has an identity card as either; reported by since 2006. We employed the following steps to the mother or verified by the enumerator. Attended assign 2008 expenditures by age group. First, we birth: child’s birth given in a public or private sector identified all centrally funded social expenditures health facility (the alternative being the respondent’s that are not contribution financed. Hence, we only or someone else’s home). Complete immunization: include treasury financed transfers to the social immunization against six main vaccine-preventable security institution that cover non-contributory illnesses: tuberculosis, diphtheria, pertussis, tetanus, programs (like the disability and old-age pension poliomyelitis, and measles, through receipt of the programs) or pension and health insurance deficits. following vaccinations before the first birthday: Green-card expenditures are included in the Ministry one dose of BCG, three doses of DPT and polio, of Health budget. As the only item outside the central and one dose of measles vaccine (BCG protects public budget, we included in our analysis the funding against tuberculosis, DTP against diphtheria, for the conditional cash transfer program and other pertussis, and tetanus). Iodine deficiency: test of social safety programs of the SYGDM (which is iodine content of the salt used for cooking indicates financed through earmarked taxes not centrally the salt is potassium iodized or contains potassium collected through the Ministry of Finance). The state iodide (at least 15 parts per million); sample of contribution to the Unemployment Insurance Fund children under five in a sample of households where is equally included in our analysis. Second, we used the iodine content of cooking salt was tested. the guidelines and explanations published by the Stunting: standardized weight-for-age below 2 Ministry of Finance (Analitik Bütçe Sýnýflandýrmasýna standard deviations of the mean of the reference �liþkin Rehber) to identify the main beneficiary group population. of the relevant expenditures. Hence, except for general health expenditures, we derived the age-beneficiary 7 The source data and method for calculating ap- relationship from the nature of the programs being proximate coverage rates for ECD programs in financed as detailed further below. Third, to arrive Turkey is as follows: Pregnancy monitoring, antenatal at per capita spending levels, we divided total care and immunizations data are calculated from the aggregate social expenditures by age group with the Turkish Demographic and Health Survey (2004). updated, age-specific population estimates provided Growth monitoring and psychosocial development by TUIK. Turkey: Expanding Opportunities for the Next Generation 33 In the health sector, we used available age-specific age groups. “Old age services”, which are also beneficiary weights from the Ministry of Health to provided by the SHÇEK, include largely for those distribute aggregate health expenditures in Turkey above the age of 65 to which group they were (Ministry of Health Turkey National Health Accounts assigned. “Family and children benefit”, which is Study 1999-2000, 2004). We also apply these same mostly covered under the SHÇEK, are assigned to weights to central government transfers to cover the age groups 0-6 years of age and 7-18 years of the deficit of the universal health insurance. age based on the total number of SHÇEK bene- ficiaries in respective age groups. “Unem-ployment For education, we apply the following age-specific benefit” (representing the government contribution classification. All pre-primary education to the Unemployment Insurance Fund) spending is expenditures are used for age group 5 and 6; primary distributed according to the share of the active education expenditures are mostly used for age working age population weights of the respective group 7-14, but a small fraction of primary education age groups. Finally, “Social Inclusion program” expenditures are counted under pre-primary spending covers the poor; immigrants; victims of expenditures because of the fact that currently crimes and violence; drug and alcohol addict. So, around 600 thousand pre-school kids are provided services under the primary schools of MONE; all expenditures of the budget are distributed according secondary education expenditures are used for age to population weights of the respective age groups. group 15-18 (items covered include General Social Protection Not Elsewhere Classified item Programs of Secondary School; Technical Secondary mainly covers transfers to SSI to cover the pension School; and other ‘Not Elsewhere Classified’). We deficit; universal health insurance (UHI) deficit and also include the salaries of preschool teachers (full- non-contributory social assistance provided by SSI time teachers) that teach at primary school facilities (including old-age pension; benefits for disabled; as benefits that accrue to the 5-6 year group. We orphans; veterans etc.). This item also covers take this amount (a total of around 400 million TL) additional payments (Ek Ödeme) to all pensioners from primary school budgets and add to the in order to cover tax refund and additional payments preprimary level. Then, we divide tertiary education to high level civil servant retirees (like ex-mayors). expenditures into two age groups 19-25 and 25-35. The treasury transfer of the UHI is covered under Public spending on college; MYOs; and master the health expenditures of the government and it is studies are classified for the age group 19-25. distributed according to the distribution of the health Expenditures of the Government for doctoral expenditures defined above. The treasury transfer programs are classified for the age group 26-35. for pension spending is distributed, using population The item “education not definable by level” consists weights, for age groups above 65 and 44-64. Social of apprenticeship training; public training centers; assistance spending is distributed according to and public vocational training. We distribute such program types and number of beneficiaries. expenditures by the approximate participant profiles Additional payments to pensioners are distributed of different training courses (from the Ministry of by using the weights of the respective ages. Lastly, National Education). Lastly, “administrative and spending outlays of the SYGDM for the conditional research activities” and “education services not cash transfers to families is divided according to their elsewhere classified” are allocated among age groups according to the actual expenditure distribution different beneficiary groups: all pregnancy and health excluding such item. aids are assigned to the 0-6 age group; education support is allocated for primary school (7-14 age) and For the overall centrally funded social protection separately for secondary school (next age bracket expenditures, we used the detailed description of using the respective population weights. Other spending the nature of the programs to distribute them across outlays of the SYDGM are classified as (i) transfer to age groups. Sickness and disability transfers, which MONE for free books for primary school kids; free are provided by the SHÇEK, are distributed in lunch provision and transportation support for relation to the population weights of the respective disabled kids, (ii) educational assistance to poor Turkey: Expanding Opportunities for the Next Generation 34 other than CCT, (iii) the other social assistance · The current relationship between a set of spending for poor. In kind transfers for primary individual and family characteristics and educa- school children is counted under social protection tional attainment is estimated using an ordered spending for age group 7-14 equivalent to the probit model. This model is used for predicting primary school age cohort Education spending other the increases in educational attainments accruing than the CCT is distributed to social protection to each individual under the examined educational 67 expenditures for age groups from pre-primary to expansion scenario. Individual heterogeneity 68 tertiary education according to their respective is taken into account to incor-porate the idea population weights. Other activities of the SYDGM that, whatever the mean expansion at the national with a poverty-focus are distributed to social level, some individuals have unobserved protection expenditures according to population characteristics that may or may not be favorable 69 weights of the respective ages. to the acquisition of education. · The observed relationships between educational 9 Distribution of pensions and other income transfers attainments, other individual and family across household members: For this calculation, characteristics and labor earnings are estimated we use the Household Budget Survey (2006) dataset using linear regression models. Labor earnings provided by TUIK. We consider transfers that are are estimated separately for formal wage earners classified under pension benefits (variable name and informal wage earners or independent emekli_yl in HBS) , social assistance transfers workers. These estimates are used for predicting (sosy_yl in HBS) and non-contributory old age the increases in labor earnings in each occupation pensions (yasli_yl in HBS). In the first step, we sum accruing to individuals who reach higher all benefits received by members of the household educational attainments. It is assumed here that at the household level. Secondly, we divide this the returns to education conditional on family total value of benefits by the number of people in and individual characteristics remain unchanged. each household and calculate the per capita benefits This assumption rests on the non-endogeneity received in each household. Finally, we aggregate of attainments as regards the unobserved deter- the benefits received by age level. Hence, we derive minants of earnings, and in the absence of general a distribution of benefits that accrue to each age category which we can then apply to aggregate equilibrium effects. Further simulations could expenditure data from the Ministry of Finance. explore alternative scenarios on the changes in the returns to education. 10 The micro-simulation is obtained using the · Fertility choices, measured by the number of methodology proposed by Bourguignon, Ferreira children per woman, are modeled using an and Lustig (2005) and Bourguignon, Ferreira and ordered probit as a function of family and Leite (2008). The micro-simulations rely on out- individual characteristics, including educational of-sample predictions based on the statistical attainments. These estimates allow predicting relationships, observed in the 2006 Turkish the change, and likely reduction, in the number household budget survey, between a set of of children chosen by each woman after the observable characteristics and the determinants of increase of her educational attainment. income and poverty, including: attained levels of · Occupational choices, measured by a categorical education, fertility choices, occupation, and labor variable indicating unemployment, formal wage earnings. employment and informal wage or independent employment, are modeled using a multinomial A model is estimated for each one of these outcomes logit as a function of family and individual and used for simulating the effects of an increase characteristics, including educational attainments in educational attainments. but also the number of children. This model of 67 This is done by translating the cut-off points of the ordered probit model, see Bourguignon, Ferreira and Leite (2008). 68 Individual residuals are drawn from the distribution assumed by the model, which is a normal distribution in the case of the ordered probit. 69 This heterogeneity is taken into account in the same way in the other equations of the model. Turkey: Expanding Opportunities for the Next Generation 35 occupational choices can be viewed as a reduced as household characteristics (household size, marriage form of a structural model of labor supply as status, number of children, etc.). potential earnings are not included among the independent variables. These estimates allow Step 2: We used this regression to calculate the us to predict the direct effect of the increase of expected probability for each woman in the survey educational attainments but also its indirect effect to be participating in the labor market (hence through fertility changes. working or looking for a job). The simulated changes in earnings, fertility and Step 3: We calculated gender specific wage income occupational choices, provide predictions of the regressions, using a two-stage estimation procedure effects of educational expansion on the structure to correct for the first-stage decision as to whether of employment, and the distribution of earnings women participate or not in the market. 70 and family income, including poverty. Per capita income depends on the occupations of each of the Step 4: Using different target rates for the female members of the household, the earnings they receive labor force participation (i.e., 29 percent as the in those occupations, and family sizes (which depend specified short-term target of the Government and 71 on the number of children). Fertility changes thus 40 percent as a longer-term target), we computed have indirect effects on welfare through occupational the number of women that would additionally enter choices and household size. the labor market, and then ‘moved’ the resulting additional women into the labor who were not The simulated educational expansion we consider participating before but who had the highest consists in an increase in the mean attainment of probability of participating according to the previously 72 individuals aged 20 to 39 years old by one year calculated probabilities (from step 1 and 2). (depending on characteristics, the educational attainment of each individual changes by a different Step 5: We estimated hypothetical incomes for these amount). This age group is chosen in order to women using the Mincer-equations and the examine the long run effects of an educational estimated parameter values from step 3. This expansion benefiting the cohorts born during the assumes that the returns to education and other two coming decades. An average increase in assets do not change albeit the labor force supply attainments by one year corresponds to the estimated increases. [Given that most of the women entering impacts of the past early child development the labor market would earn relatively modest programs in Turkey. However alternative scenarios incomes, such assumption would be justifiable of educational expansion can be explored. given that the labor demand curve for relatively low incomes is highly elastic {World Bank 2009a}]. 11 We used a simple model of labor force participation and income earnings regressions to assess the possible Step 6: Lastly, we calculated the change in household consumption and poverty impacts of an increase in income and consumption, assuming that the female labor force participation in Turkey. additional household income would increase actual consumption in full. We are hence able to assess Step 1: Using the Household Budget Survey for 2006, (i) the change in overall income of all households for all women 15 and above, we estimated a probit (which would provide a lower bound estimate – regression of labor force participation, controlling for since it does not take multiplier effects into account); all available individual (education, age, etc.) as well and (ii) changes in the poverty rate. 70 The poverty rate is computed using per capita income (instead of consumption) and the poverty line used by TUIK. 71 Household non labor income is assumed to remain unchanged. 72 A variable for educational attainment in completed years is constructed using the categorical variable for completed levels of education available in the HBS 2006 survey. Note: Note: Note: