Table 8.

Estimating the reduction in stunting from increasing various education policies in high-burden countries in South Asia and sub-Saharan Africa

	Alternative scenarios
	Universal primary school completion (6 years)	Universal middle school completion (9 years)	Universal upper secondary completion (12 years)
Reduction from both sexes	−2.46%	−6.04%	−10.32%
Reduction from girls' education	−1.45%	−3.90%	−6.23%
Reduction from boys education	−1.01%	−2.15%	−4.10%

Source: Authors’ estimates from DHS data. See text for details.

Notes: These estimates are based on three steps. First, we separately estimated regressions of stunting against education levels for high-undernutrition burden countries in South Asia and sub-Saharan Africa, analogous to the HAZ regressions in Table 5. Second, we estimate mean levels of education by bracket for South Asian and sub-Saharan African DHS surveys post-2005 and parents aged 20–29 years of age (i.e. the most recent generation of parents). Third, we estimate counterfactual distributions of education for the current generation: universal primary parents with 0–6 years of education are now assigned 7–9 years; universal secondary, where parents with 0–9 years of education are now assigned 10–12 years of education; universal tertiary education, where the whole population is assigned 13-plus years of education. Fourth, we calculate predicted changes in stunting rates from these counterfactual distributions and the estimated coefficients derived from step 1. In each scenario we always hold the proportion of parents from higher education levels constant (e.g. in scenario 1 we reallocate parents in the 0–3 and 4–6 year brackets to the 7–9 year bracket, but keep the 10–12 and 13 + proportions the same). Note, also, that since the number of children 0–14 in South Asia and sub-Saharan Africa is relatively similar, we equally weight the results from each region to derive an aggregate for both regions.