Skip to content

The Economist's modelling of nutritional outcomes, informed in part by demographic projections

Notifications You must be signed in to change notification settings

TheEconomist/nutrition

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

33 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Demography, nutrition and cognition

This repository houses The Economist's data and scenarios of nutritional outcomes based on past and present incidence of nutritional deficiencies and demographic projections.

Contents:

  1. Estimates and projections to 2050 of stunting in newborn children and anaemia in pregnant women
  2. Estimates of proportion of kids who are developmentally delayed, by country
  3. Charts and stats based on these metrics

Data:

Estimates are all available in the "output-data" folder. This folder contains:

Analysis

The analysis is organized into four scripts.

01_stunting.R - This script takes as its inputs UN population projections and WHO country-year estimates of stunting prevalence among children under five years of age. Stunting is defined as abnormally low height for age. Where stunting rates are missing, which is the case mainly in a few high-income countries or tiny countries, stunting rates are assumed to be equal to the UN sub-region stunting rate (e.g. "Western Europe").

This script assumes that a child's probability of being stunted is equal to the rate of stunting among those under five in the country in the child's year of birth. This enables us to calculate the world-wide stunting rate over time. It then calculates three scenarios. The first is if stunting continues at current rates, in this case 2022 values (which are the latest data point). The second is what would happen if trends of the past decade continue, or to be precise, the most recent decade of data. To estimate this, it uses a quasi-binomial generalised linear model with logit link with country fixed effects interacted with a year trend. The script offers various calibration plots to assess if such projections are reasonable. The third scenario estimates what would happen if rates declined twice as fast as projected based on trends of the past decade (and no country did worse).

Stunting rates

02_anaemia.R - This script takes as its inputs UN population projections and WHO country-year estimates of anaemia prevalence in pregnant women. The approach is largely identical to that of the script to investigate stunting rates. Where rates are missing, which is the case mainly in a few high-income countries or tiny countries, rates are assumed to be equal to the UN sub-region rate (e.g. "Western Europe").

Like above, the script then calculates three scenarios. The first is if anaemia in pregnant women continues to occur at current rates, in this case 2019 values (which are the latest data point). The second is what would happen if trends of the most recent decade continue, or to be precise, the most recent decade of data. To estimate this, it uses a quasi-binomial generalised linear model with logit link with country fixed effects interacted with a year trend. The script offers various calibration plots to assess if such projections are reasonable. The third scenario estimates what would happen if rates declined twice as fast as projected based on trends of the past decade (and no country did worse).

Anaemia rates

03_general_development.R - This script combines the stunting, anaemia, and demographic data with data on the proportion reaching developmental milestones from surveys, specifically, "children aged 36-59 months who are developmentally on track in at least three of the following domains: literacy-numeracy, physical development, social-emotional development and learning", as well as GDP per capita at purchasing power parity from the World Bank. Each survey is combined with the stunting, anaemia, birth and gdp per capita estimates from that year in the country in question.

Developmental milestones

04_world-totals.R - This script loads data on worldwide estimates and scenarios of stunting and anaemia rates and absolute numbers so that one can quickly inspect and chart it.

Many of the charts used based on this data have versions available in the plots folder of this repository. If you have any questions about this analysis or spot an error or opportunity for improvement, please open an issue or email [email protected].

Notes

  • Our analysis is based on data and estimates which have considerable uncertainty. This should be kept in mind when interpreting the figures. While uncertainty in global figures should be less than the sum of uncertainty in country-level data, we are not able to provide it, as the micro-data required to quantify it is not available (it depends on correlations in errors between countries).
  • Our scenarios are just that, scenarios: they should not be seen as predictions.
  • Data on malnutrition is not perfect. Some countries have no data. In these instances we have had to rely on sub-regional averages. Data is also released with a lag: for anaemia in pregant women, data ends in 2019, and for stunting rates, in 2022.
  • To estimate worldwide rates, country-level estimates were weighted by number of births in the country. While this is, as far as we know, the best approach possible, the number of pregnant women is slightly different from the number of births in a country in a given year. If these differences varies in a way correlating with anaemia rates, they could affect our estimates. Relatedly, stunting rates in under-5s is taken as a proxy for the expected stunting rate among children born in the country in that year, a highly related but different quantity. It may tempting to think that, since stunting rates falling in most countries, stunting rates in under-5s should be higher than those of newborns. However, approximately 1 in 27 children die before the age of 5, and these are disproportionally the malnourished, pushing the rate up.

Sources:

Suggested citation

The Economist and Solstad, Sondre (corresponding author), 2024. Malnutrition data and projections.

About

The Economist's modelling of nutritional outcomes, informed in part by demographic projections

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published