Micro-level Estimation of Child Malnutrition Indicators and Its Application in Cambodia
One of the major limitations in addressing child malnutrition is lack of information that could be used to target resources. By combining demographic and health survey (DHS) and population census data, the author disaggregates the estimates of the prevalence of child malnutrition in Cambodia from cu...
| Main Author: | |
|---|---|
| Language: | English |
| Published: |
World Bank, Washington, DC
2012
|
| Subjects: | |
| Online Access: | http://documents.worldbank.org/curated/en/2005/07/6054363/micro-level-estimation-child-malnutrition-indicators-application-cambodia http://hdl.handle.net/10986/8206 |
| Summary: | One of the major limitations in addressing child malnutrition is lack of information that could be used to target resources. By combining demographic and health survey (DHS) and population census data, the author disaggregates the estimates of the prevalence of child malnutrition in Cambodia from currently available 17 DHS strata into 1,594 communes. The methodology is built on the small-area estimation technique developed by Elbers, Lanjouw, and Lanjouw. The author extends it to jointly estimate multiple indicators and to allow for a richer structure of error terms. Average standard errors for the commune-level estimates in this study were about 4 percent, a magnitude comparable to those for stratum-level estimates derived from DHS only. The author demonstrates three applications of these estimates. First, he explores the relationship between malnutrition, consumption poverty, and inequality. The nonlinear effects of consumption on nutritional status of children are a key component of the relationship. Second, he conducts a decomposition analysis of health inequality and finds that the between-location share of health inequality is lower than with consumption inequality. Finally, he evaluates the potential gains from geographic targeting. The author finds that the savings in the cost of a nutrition program from commune-level targeting is on average at least two to three times higher than that from stratum-level targeting when the per capita cost of the program is fixed. |
|---|