Household Poverty in Rural Côte d'Ivoire

• In Event: Developments and challenges in economics and finance of education: Evidence from LMICs and high-income countries

Proposal

Research overwhelmingly testifies that poverty is a significant barrier to children’s healthy development (Engle & Black, 2008; Shonkoff et al., 2012; Rakesh et al., 2021). Poverty, particularly chronic poverty, is a significant stressor that shapes children’s lives and leads to opportunity gaps that threaten children’s development (Luthar et al., 2000; Duncan et al., 2010). This paper presentation explores household poverty and children’s development from a sample of 1,847 households from 140 villages in rural Côte d’Ivoire (from Aboisso, Méagui, and Bouaflé regions) that belong to economically vulnerable cocoa-farming communities. Data includes survey results from 1,358 children (Boys n = 705, Girls n = 643; Ages 5-15 years (M = 9.07, SD = 2.10). Through this paper presentation, I investigate children’s education (literacy, numeracy, and socioemotional scores) and health outcomes (weight for age and height for age) and their associations with household poverty using two measures. The two measures are (1) the Multidimensional Poverty Index (MPI), a weighted-sum index of household poverty, and (2) a “neighborhood access” index developed by the author (Alkire & Foster, 2012). Geo-coded data extracted from OpenStreetMap is used to create a "neighborhood access" variable by the author via principal component analysis, including factors like access to hospitals, schools, roads, electrical grids, and violence (Brusilovskiy & Salzer, 2012). This measure, along with the MPI, is used to explore children's educational and health associations with separate OLS regression models in R studio.

The MPI considers three poverty dimensions to capture household poverty: health and food insecurity, education, and standard of living. Using the poverty thresholds defined by the Alkire-Foster method, over half of the study sample falls under the “severely multidimensionally poor” with an MPI score >= 0.5. Only 14.3% of the sample is not multidimensionally poor. Using unconditional OLS regression models, higher household poverty using the MPI was associated with lower literacy and numeracy scores at non-significant levels. Socioemotional scores ( = -0.20), and child health ( = -1.08) were negatively associated with household poverty at the 0.05 level. The “neighborhood access” measure was associated with lower numeracy scores ( = -0.15) at the 0.05 level. Other child domains were non-significant. These results highlight the robust associations with household poverty and children’s cognitive and physical development, shedding light on the importance of policy focusing on poverty alleviation. Moreover, the difference in associations between the MPI and the neighborhood access measure show that types of poverty may have varying results for children, which can help shape policy. This work extends a previously Western-focused topic to rural West Africa, providing broader insights into the use of novel measures in education and education-adjacent research.

Authors

• Anahita Kumar, GSE, University of Pennsylvania

• Sharon Wolf, University of Pennsylvania