SES and growth in stature during early adulthood in low- and middle-income countries

This article uses data from many low- and middle-income countries (LMICs) to provide systematic support for the hypothesis that in low-resource settings adult height is achieved later in life, leading to a reduction in the height gap that usually exists between individuals from low and high socioeconomic status (SES). I first use height measurements for more than 1.4 million adolescents and young adults aged 15-29 (mostly females) from 180 Demographic and Health Surveys (DHS) conducted in 63 countries between 1991 and 2020. Higher per capita GDP - a coarse proxy for overall economic and epidemiological conditions - predicts smaller differences in height between cohorts aged 20-24 and those aged 15-19. Among females, the slope is -0.41 centimeters (s.e. 0.08, p-value <0.01), while for boys it is -1.04 (s.e. 0.35, p-value <0.01). Within-country estimates show that different proxies of high SES predict, on average, smaller differences in height between older and younger cohorts. DHS data are cross-sectional, and do not allow the estimation of individual growth curves, which are proxied by average differences in stature between adjacent cohorts of individuals surveyed at the same time, and I discuss the drawbacks of this approach. Longitudinal data from the Young Lives study allow the calculation of individual-level growth in stature, although only from four countries. Again, SES is inversely correlated with height growth in post-adolescence. When studying adult height in LMICs, researchers should measure height after age 20, test age-threshold robustness, and be aware that SES gaps at 18-19 may be overstated.