Are Children in West Bengal Shorter Than Children in Bangladesh?
Total Page:16
File Type:pdf, Size:1020Kb
COMMENTARY economically and statistically signifi - Are Children in West Bengal cantly shorter than children in Bangla- desh, on average. Many factors may con- Shorter Than Children tribute to this gap, and a full accounting is beyond the scope of this note. Never- in Bangladesh? theless, we document that differences in the disease environment – namely, that children in West Bengal are exposed to Arabinda Ghosh, Aashish Gupta, Dean Spears much more open defecation than chil- dren in Bangladesh – can fully statisti- Children in West Bengal and lthough researchers have long cally account for West Bengal’s height Bangladesh are presumed to established that height differences defi cit, especially in combination with differences in women’s status. share the same distribution of A between populations are largely driven by environmental differences – genetic height potential. In West especially net nutrition and disease Initial Comparisons Bengal they are richer, on average, (Bozzoli et al 2009; Coffey 2013; Hatton Table 1 (p 22) offers a simple comparison and are therefore slightly taller. 2013) – some commentators have recently of human development summary statis- asked whether the exceptional height tics in West Bengal and Bangladesh. However, when wealth is held defi cit of India’s children may be merely Panel A collects aggregate statistics that constant, children in Bangladesh genetic (Panagariya 2013). The answer we report from published sources. We are taller. This gap can be fully is of policy importance because adult follow the format of Table 3.2 of Sen and accounted for by differences in height refl ects health and net nutrition Drèze (2013: 51), which we extend to in the critical fi rst few years of a child’s West Bengal. Panel B reports our own open defecation, and especially by life. Lifelong physical and cognitive computations from the data used in this open defecation in combination development are lastingly shaped dur- paper’s analysis. We pool data from with differences in women’s ing this early period (Case and Paxson India’s 2005-06 National Family Health status and maternal nutrition. 2010; Spears 2012), and the average Survey (NFHS) and Bangladesh’s 2007 height of a population is a key indicator Demographic and Health Survey (DHS). of the health and economic productivity As part of an international DHS project, of the adult workforce (Case and Paxson these surveys are designed to be compa- 2008; Vogl forthcoming). India cannot rable and can be used in combination. afford to misunderstand the causes of Summary statistics are computed to ref- child height. lect the data in our analysis; thus chil- To illustrate these facts about popula- dren under 5 are the observations, not tion height, in this note, we focus on a households. If young children are dis- simple comparison between children proportionately concentrated in poorer living in the Indian state of West Bengal or otherwise disadvantaged households, and children living in the neighbouring then these fi gures will suggest a lower country of Bangladesh. These societies level of human development than were split into different political states nationally representative fi gures would. recently on a genetic time scale and Two basic, well-known facts emerge much migration continues. Cultural, from these summary statistics. First, geographic, and agricultural factors are people in West Bengal are richer than shared between these populations, people in Bangladesh – they have a higher although Bangladeshis are more likely income per capita and are more likely to be Muslim. to own private assets such as radios, To answer our question simply – no, bicycles, motorcycles, and telephones Our conclusions are our own and do not children in West Bengal are taller, on av- (although slightly less likely to have necessarily represent the views of any organisation. erage, than children in Bangladesh. electricity). Second, Bangladeshis have However, families in West Bengal are advantages along other dimensions of Arabinda Ghosh is in the Indian Administrative also richer. Food and other care that human development – more women can Service, West Bengal; Aashish Gupta is with the Research Institute for Compassionate money can buy are important determi- read, mothers weigh more, women are Economics, Sitapur, Uttar Pradesh; and Dean nants of early life net nutrition, and more likely to participate in the eco- Spears ([email protected]) is with the therefore of attained height. We show nomy and politics, and a much smaller Centre for Development Economics, Delhi that at the same level of socio-economic fraction of the population defecates School of Economics, and RICE. status (SES), children in West Bengal are in the open without using a toilet Economic & Political Weekly EPW febrUARY 22, 2014 vol xlIX no 8 21 COMMENTARY or latrine.1 One exception to this trend born by Caesarean sec- Table 1: Summary Statistics is infant mortality, which is lower in tion; and the mother’s Bangladesh West Bengal India West Bengal; this is consistent with the age when the child was Panel A: Published aggregate statistics Population (millions), 2011 142 91 1,241 multidimensionality of health (Coffey born. In some specifi ca- GDP per capita (PPP), 2011 1,569 2,586 3,203 et al 2013a). tions, we also control for Population density, 2011 964 1,029 382 Using the same data, Figure 1 plots the a child’s mother’s height; Urban population (%), 2011 29 32 32 average height-for-age of children under although we primarily Infant mortality rate, 2011 37 26 44 5 at each month of growth and develop- intend this variable as a Open defecation, 2011 (%) 7.0 38.6 49.8 SES ment. The negative numbers on the ver- further marker of Female labour force participation, 2010 (%) 57 18 26 tical axis indicate that children from (refl ecting the mother’s Women’s share of legislative seats, 2011 (%) 18.6 11.2 10.7 both countries are shorter than would own upbringing), control- Panel B: Our computations from DHS data, representative of children under 5 be average for a population of healthy ling for it should also re- Height-for-age z-score -1.75 -1.70 children. The fi gure displays a familiar move any fi nal doubt of Household open defecation 0.21 0.53 pattern – increasing height shortfalls as the genetic comparability Local (PSU) open defecation 0.20 0.52 growth defi cits accumulate until about of these populations. Mother literate 0.61 0.54 two years of age, at which point growth Finally, we individually Mother’s height (cm) 150.4 150.6 paths are largely determined and adult add three specifi c controls Mother’s BMI 20.1 19.3 stunting is likely. Bangladeshi children for factors known to be Urban residence 0.206 0.209 Has electricity 0.439 0.422 fall further below the norm than chil- important for child height Has radio 0.232 0.284 dren in West Bengal do; so, without ac- in a simple attempt to ac- Has refrigerator 0.064 0.064 counting for differences in their materi- count for the height gap – Has bicycle 0.223 0.624 al environments, it is clear that children exposure to open defe- Has motorcycle 0.033 0.070 in West Bengal are taller, on a verage. cation; mother’s literacy; Has telephone 0.012 0.051 and mother’s body mass SES index used in Figure 2 -0.166 0.251 Accounting for Wealth index (BMI) as an indica- Figures in Panel A are indicative rather than definitive as they are based on various How would these results differ if we did tor of maternal social sources of survey and census data, not all originally constructed to be comparable. Source: All figures from World Development Indicators (World Bank, 2013) adjust for differences in wealth? For the status, a predictor of in unless otherwise stated. Population for India and West Bengal from Government main analysis of this article, we use the utero nutrition, and a of India (2011a), Statement 3, p 47, for Bangladesh from Bangladesh Bureau of Statistics (2011a); GDP per capita (PPP) 2011 for West Bengal calculated using pooled DHS data to estimate descriptive correlate of breastfeed- nominal state domestic product figures from Government of India (2013a), p A13; regressions of the form ing quality (Coffey 2013). official exchange rate (local currency unit per $) and PPP conversion factor from World Bank (2013); population density (persons per sq km) for Bangladesh from heightip=β0+β1 West Bengalp+β2mother’s Figure 2 (p 23) presents Bangladesh Bureau of Statistics (2011), for India and West Bengal from Government height + SES θ + γcontrols + ε , a non-parametric sum- of India (2011a); urban population for India and West Bengal from Government of ip ip ip ip India (2011b); infant mortality rate for India and West Bengal from Government of where i indexes individual children u nder mary of our fi rst result – India (2012), proportion of population defecating in the open for Bangladesh from fi ve years old, and p represents local plac- at all levels of a socio- Bangladesh Bureau of Statistics (2011b), for India and West Bengal from Office of the Registrar General and Census Commissioner (2012), female labour force es, in this case survey primary sampling economic status index, participation for India and West Bengal from Statement 24, Government of India units (PSUs), according to which we cluster children in West Bengal (2013b); women’s share of legislative seats from Sen and Drèze (2013). 2 standard errors. Estimates are weighted are shorter, on average. Figure 1: Unconditionally, Children in West Bengal Are Taller according to DHS sampling weights. Height In other words, although -.5 is the height-for-age z-score of a child children in West Bengal and West B engal is an indicator that the are taller overall, they are child is from the West Bengal sample.