Policy Brief #2:
Child poverty in Nepal

1. What did the recent NLSS (2022/23) reveal?

   • At the household level estimated poverty rate of \(20.22\%\) at new NLSS poverty line of NPR 72,908 per capita per year. (NSO 2024)

2. Larger households and families with more children below 6 are significantly at higher risk of poverty.

3. Lost opportunity to reveal any information about children living in poverty. Child poverty poses a serious threat to children [Engle and Black (2008)](Lancet 2023) as well as Nepal’s vision of an inclusive, equitable society with improved human capital and productivity (16th Plan approach paper).

4. Urgency to fill the knowledge gap regarding child poverty as SDG 1 calls for an end to poverty in all its forms including child poverty.

1. Using machine learning applied to UNICEF CFT data (Aug 2023) to estimate household monthly incomes from reported household income. Estimate per capita income by using demographic structure of households.
2. Estimate the percent of children who are poor and the percent of poor who are children. Examine other multidimensional challenges faced by children in poor and non-poor households.
3. Simulate the impact of making the U5CG universal on per capita income distribution and poverty.

This presentation is about children and the multifaceted challenges they face including income or money. The rationale is based on the paucity of research in this area. NLSS 2022/23 did not highlight the situation of children directly and neither did the MPI report, although the latter inherently uses children’s indicators in its construct. This study utilizes the reported income data from CFT 11 and leverages the power of machine learning, both supervised and unsupervised, to decipher some of the monetary and other challenges that children in Nepal face. Starting with their families, the analysis then shifts to children themselves and examines disparities among them in terms of their HH income, place of residence and province, disability status, composition of the family, whether the family is struggling for food, whether HH has health insurance, gender, age and a range of other background characteristics. Goal is to derive income-poverty among children. What percent of children are poor? What percent of the poor are children? What are the other multi-sectoral challenges children face, and how are these correlated to income and other lenses. Finally, we imagine what would happen if the child grant were made universal at different levels of benefits.

Linear models and the LASSO models were the most divergent. Can see some convergence among other models. By comparing reported and predicted HH monthly incomes, we get an idea of under-reporting. However, this is not the point of this study. We use predicted Hh incomes (from random forest model) to generate per capita incomes and poverty. The average household income was estimated at NPR \(50,750.25\).

\(26\%\)OF FAMILIES WITH CHILDREN ARE ESTIMATED TO BE LIVING IN INCOME POVERTY. BUT THERE IS A WIDE VARIATION BY PROVINCE. IN THREE PROVINCES POVERTY RATES ARE HIGHER THAN \(30\%\) AND HIGHER THAN \(35\%\) in 2.

This study demonstrates that households with children, particularly those in the lower income quintiles, frequently turn to debt or asset sales to manage financial strains, alongside grappling with various challenges such as accessing nutritious food.

The data show some evidence of ‘age heaping’ (rounding off ages). There is a gender gap between boys and girls that needs further investigation.

Among all household members, those who are in the age group 12-20 years, face the highest risks of being in a household that is below the poverty line. There is a statistically significant difference between boys and girls. This is an important finding as it underscores the need to focus on gender dimensions and this age group in addition to those under the age of five. Due to demographic changes, the age profile of poverty will increasingly consist of children who are older.

"Poverty exacts a devastating toll on children, impeding physical growth, hindering social and emotional development, and shortening life expectancy. It inflates infant and child mortality rates and raises the likelihood of chronic health conditions, ravaging children with preventable diseases. Moreover, poverty perpetuates achievement gaps, amplifies parent stress, and undermines parenting practices. The dire consequences of poverty extend to children's daily lives, introducing hunger, neglect, insecurity, and instability. Additionally, poverty increases the prevalence of violence experienced by children directly and witnessed within their communities, compounding the already profound challenges they face."

(Lancet 2023)

HH level risks and challenges translate directly into challenges and risks for children. In addition to the risk of living in households below the poverty line, children also face other household risks: \(35\%\) of children are residing in households struggling for food. \(64\%\) of children were living in households without access to health insurance. \(12\%\) of children were affected by DRR hazards. Children are also affected by coping mechanisms deployed by families. Nearly half of all children (\(45\%\)) live in households that are selling households assets to cope financially and \(30\%\) are living in households that are incurring debt to cope. The disparity between children living in poor households and those that are not is striking.

1. This study has used machine learning and unveiled significant findings regarding child poverty and the various multidimensional challenges children face. The data indicate that children are exposed to numerous risks stemming from household income and other vulnerabilities their families might encounter.

2. The analysis underscores the urgent need to accelerate progress toward Sustainable Development Goal 1—ending poverty in all its forms, including child poverty.

3. To achieve this, a comprehensive approach is necessary, one that addresses both income poverty and other facets of child deprivation in Nepal. This approach must effectively address disparities linked to geography, caste/ethnicity, disability status, and the nutritional and health status of households, among other factors.

4. Our observations highlight actual outcomes, not just potential scenarios or what could be. One promising avenue in this context is the Under-5 Child Grant (U5CG). While it alone cannot resolve the broader issues of children’s agency and rights, it holds significant potential for alleviating income poverty, as evidenced by this study.