The Response of Child Nutrition to Changes in Income.pdf

The Response of Child Nutrition to Changes in Income: Linking Biology with Economics* Harold Alderman y y World Bank 1818 H St., Washington, DC 20433, USA. e-mail: Abstract It is regularly pointed out that despite impressive economic growth India has the largest number of malnourished children in the world. It also has the largest number of people with diabetes despite moderate levels of obesity. These two observations may in fact be linked; p
of 18
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  The Response of Child Nutrition to Changes in Income:Linking Biology with Economics* Harold Alderman y y World Bank 1818 H St., Washington, DC 20433, USA. e-mail: Abstract It is regularly pointed out that despite impressive economic growth India has the largestnumber of malnourished children in the world. It also has the largest number of people withdiabetes despite moderate levels of obesity. These two observations may in fact be linked;poverty breeds malnutrition in children and chronic diseases in adults. This article reviewsevidence on the critical role of early nutrition both for long-term health and also for reducingthe intergenerational transmission of poverty. (JEL codes: I12, I14, O15) Keywords:  nutrition, child development 1 Introduction There is likely little doubt on the validity of either the proposition thatincome growth reduces malnutrition or that reducing malnutritionpromotes income growth. How much emphasis is placed on the latter,however, depends, in part, on the expectations on the magnitude of theformer. In particular, if there is a sharp gradient of nutritional status asone moves away from poverty then a different set of programs might beprioritized than if—as evidence in Haddad et al. (2003) implies—incomegrowth even when evenly distributed over a population has a relativelymodest, albeit significantly positive, impact on undernutrition rates.Because it is comparatively easy to derive quasi-reduced form estimatesof the association of child malnutrition and current household income thisrelationship can be used to both guide development priorities as well asaddress pragmatic operational questions such as how much transfer isneeded to effect a meaningful change in health-seeking behavior(Fernald et al. 2008). This article therefore begins with a discussion of two observations on the relation of income and income growth to nutri-tional status. One observation is on the magnitude of the association of income and nutrition and the other focuses on evidence on the timing of income. The next section explores new evidence on how these economicpaths may relate to aspects of physical growth again choosing to highlighttwo particular features: (i) the timing of critical periods of growth and (ii)some recent studies that potentially have a bearing on why obesity and * Paper prepared for CESifo workshop on Malnutrition in South Asia Venice InternationalUniversity, San Servolo, Venice 20–21 July 2011.   The Author 2012. Published by Oxford University Presson behalf of Ifo Institute, Munich. All rights reserved.For permissions, please email:  256 CESifo Economic Studies, Vol. 58, 2/2012, 256–273 doi:10.1093/cesifo/ifs012Advance Access publication 2 May 2012   b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on S  e  p t   e m b  e r 2  7  ,2  0 1  3 h  t   t   p :  /   /   c  e  s i  f   o . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om   malnutrition are often observed in the same households, never mind thesame communities. The final section links these themes. 2 Some observation on the path from income growth to childgrowth A. The global evidence that income growth is not a sufficient condition forthe rapid reduction of malnutrition On average, country-level rates of undernutrition—defined as low weightsfor age—decline at roughly 50% of the rate that gross national products(GNP) per capita increase. This cross country evidence differs little fromresults generally derived from household surveys at the individual level(Haddad et al. 2003). Likewise, anemia—defined as hemoglobin levelsbelow 10.9g/dl—declines at roughly 25% of the rate of income growth(Alderman and Linnemayr 2009).The majority of studies reporting such evidences are based oncross-sectional surveys and, as always, there are challenges to the causalinterpretation of these associations. However, since anthropometricsurveys are increasingly available, one main issue for deriving such esti-mates at the individual level for any population or subpopulation strati-fied by classifications such as gender or age is the source of expenditure orincome data. Many data sets with information on anthropometric meas-urements or on anemia such as Demographic and Health Surveys [DHS]lack income or expenditure information. Filmer and Pritchett (2001) have,however, shown that asset indices can be conveniently used to map therelationship of economic status and nutrition. This technique—nowwidely employed—also addresses the fact that a graphic or regressionalanalysis of the association of an outcome with one that is measured withsubstantial error (as are expenditures and income) is biased towards zeroin proportion to the measurement error. The construction of an assetindex sweeps out most measurement error and thus reduces this biasappreciably.Alternatively, a measure of income or expenditures that is predictedfrom assets also has this appreciable advantage over directly reportedincome. This treatment for errors in variables is closely related to theuse of instrumental variables to address the simultaneity or endogenityof income, although the issue of exclusion restrictions does not necessarilyarise when considering errors in variable bias. In any case, the mismeas-urement of income or expenditures is plausibly the dominant issue relativeto the lagged endogeneity of assets. That is, the former is known to beappreciable while the latter is unknown. While very young children do notdirectly contribute to asset accumulation, other possible sources of bias CESifo Economic Studies, 58, 2/2012  257 Response of Child Nutrition to Changes in Income  may reflect either the impact of a child’s illness on parents’ ability toacquire assets or the correlation of unobserved preferences for savingsover current consumption and health seeking behavior. 1 An asset index provides only an ordinal proxy for income or wealth.Often the index is reported in quintiles as illustrated in Table 1 or similaraggregations. Such tables confirm that malnutrition rates are high evenamong those who are not poor. For example, using the data in Table 1,where all households in the lowest two quintiles by wealth in Pakistan tohave the characteristics of the middle quintile, poverty in the countrywould be virtually eliminated, yet over 38% of the entire population of children would still be malnourished.Such an association of malnutrition and asset rankings can be used toderiveanapproximationofincomeelasticitiesiftherankingofassetscanbeassociated with the average income of each quintile using alternative data.Similarly, it is straightforward to construct average elasticities of income Table 1  Nutrition and poverty: prevalence of child underweight by wealthquintilesRegion Country Lowest 2nd 3rd 4th Highest South Asia Bangladesh 59 53 45 43 30India 61 54 49 39 26Pakistan 54 47 43 37 26 Africa Benin 29 30 23 20 10Burkina Faso 42 40 41 39 22Ethiopia 49 51 51 45 37Mozambique 31 28 26 19 9Rwanda 27 30 28 24 14Tanzania 25 26 22 20 12Uganda 27 26 25 19 12 Source : Gwatkin et al., Country Reports on HNP and Poverty: Socio-economic Differencesin Health, Nutrition and Population, April 2007. Note : Data for children <5 years. Below – 2SD. 1 Testing the degree to which unobservable preferences might bias an income response inthis context would require creativity. The standard approach to instrumental variables orthe implementation of a randomized control trial might provide some insights but theseare also subject to the question of whether the local response represents the use of tran-sitory or permanent income. Theoretically, one would expect more investment off theformer. However, as food is both a consumption good as well as a key input into childnutrition, the expected relative magnitudes are hard to bracket. 258  CESifo Economic Studies, 58, 2/2012H. Alderman  growth at anaggregate level. For example, in the1990s, Indiaexperienced agrowth rate in GDP per capita of 5.3% whereas malnutrition declined by1.5% (World Bank 2006). The implied average elasticity of    0.28 is lowerthan the rule of thumb of    0.5 mentioned above and lower than that of China (  0.67) or Bangladesh (  0.73) over the same period.Indeed, results from India on the prospects of addressing malnutritionprimarily through income growth are generally discouraging. For exam-ple, Subramanyam et al. (2011) find that state-level economic growth inIndia offers little or no explanation for reductions in malnutrition between1992 and 2006. Since this analysis of three rounds of the National FamilyHealth Survey also includes family wealth among the covariates and sincethere were strong gradients of the risk of underweight with poverty as wellas maternal and paternal education in this data, the result hints at the roleof state infrastructure and execution of its programs rather than privateresources. This does not completely remove the puzzle of the relatively lowresponse to growth, but it does help locate it.Still, for many purposes a cardinal estimate of the impact of incomegrowth is required at a more disaggregated level. Even in the absence of direct survey data on income or expenditures it is nevertheless often pos-sible to map expenditures to the same assets that are commonly used toconstruct an asset index using data from an income and expendituresurvey and then use these parameters to predict expenditures within aDHS or similar survey that has information on assets. Such an approachwas used within the context of a randomized longitudinal trial of a com-munity nutrition program in rural Uganda. This estimate found that whileincome growth will have a positive and statistically significant impact onundernutrition, a 5% rate of per capita income growth—substantiallylarger than the average for the decade that preceded the study—wouldtake 33 years to reduce underweight rates by half (Alderman 2007).As mentioned, such indications of the magnitude of improvements givenincome growth can be useful in policy projections. Such a starting pointmay,however,notfullyguidethedesignofcashtransferssupportprogramssince there is some evidence that the source of income matters, eitherthrough labeling effects (Kooreman 2000; Paxson and Schady 2010) or through gender control. Ideally, to understand the path from income tonutrition one would like to complement a reduced form approach with anutrition production function to gain insights on the contribution of keyinputs in overall health as well as assess how income and prices affect thedemand for these inputs. While health production functions are likely to bemore stable than reduced form estimates (Rosenzweig and Schultz 1988),the difficulty in finding suitable identifying restrictions for input demandremains an obstacle to their application. Randomized controlled trials haveserved to derive the impact of selective inputs, particularly micronutrients, CESifo Economic Studies, 58, 2/2012  259 Response of Child Nutrition to Changes in Income

Online marketing

Jul 24, 2017

Green Architecture

Jul 24, 2017
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks