ANNEX 8: Methodology used for the subnational estimation of cost and affordability of a healthy diet using household survey data for selected countries in Africa in Chapter 4

The cost and affordability of a healthy diet in selected countries in Africa were estimated across URCAs applying the FAO Healthy Diet Basket (HDB) methodology, which comprises six food groups.bl However, results are not comparable with the global CoAHD indicators presented in Chapter 2 (see Box A8.1). Food prices and income distributions were obtained from 11 household consumption and expenditure surveys conducted between 2018 and 2019 (Table A5.1).

Box A8.1 Methodology – Global and subnational estimation of the CoAHD

The estimation of the global monitoring indicator of the cost and affordability of a healthy diet (CoAHD) (Chapter 2) and the subnational estimation by URCA in Section 4.2 follow the same methodology. However, results are not comparable for three main reasons:

  • Food item prices. In the global monitoring, prices from the World Bank International Comparison Program (ICP) are used, whereas prices used in the analysis presented in Chapter 4 are computed from household surveys.
  • Income distribution. In the global monitoring, the affordability indicator is computed using the estimated income distribution in a given country from the World Bank’s Poverty and Inequality Platform (PIP); whereas total household expenditure used in the analysis presented in Chapter 4 is computed from household survey data to estimate its distribution as a proxy for income distribution.
  • Percentage of income that can be credibly reserved for food. In the global monitoring, this percentage is set equal to 52 percent – that is, the average percentage of income spent on food in low-income countries based on the national account expenditure data from the World Bank ICP. In the analysis of Chapter 4, on the other hand, the average food expenditure shares of households belonging to the lowest expenditure quintile in each URCA are applied.

The analysis took place in four stages. In the first stage, the household consumption and expenditure survey data were georeferenced using the URCA dataset. In the second stage, the prices of food items were derived from household food expenditure modules reporting the quantity bought and amount spent by households based on seven-day recall.bm Values were reported for specific food items, thus allowing for computing of the revealed price (i.e. the unit cost) by food item. The price for each food item was obtained as a geometric meanbn of the revealed prices in each URCA of each country. Note that food items not reported in a specific spatial unit were considered as not available in that area.

In the third stage, the food items for the subnational (i.e. at the URCA level) HDB were selected. The HDB composition was fixed in terms of daily caloric contribution of the six food groups, as per the HDB of the global CoAHD monitoring indicators, but the specific food items in the HDB were allowed to change across URCAs. More specifically, the least-cost item in each food group was selected in each URCA for each country. In this way, the composition of the HDB accounts for spatial variation in terms of prices and availability, as well as reflects items consumed by the population in each URCA.bo To compute the cost of an item needed to meet the HDB caloric requirement, prices (as described above) and the nutrient conversion table developed for each survey (based mainly on the FAO/INFOODS Food Composition Table for Western Africa [2019]) were used.

In the final stage, the measure of affordability of a healthy diet was obtained by comparing the daily cost of the HDB with the daily per capita household income available for food. Total household expenditure, including value for own production, was used as a proxy for income. The share of expenditure that can be credibly reserved for food was set equal to the average food expenditure share of households belonging to the lowest quintile of the income distribution of each URCA. The choice i) aligns with the global CoAHD indicator methodology where the average food expenditure share of low-income countries is adopted, and ii) takes into account different levels of economic development across the rural–urban continuum.

When summary results are presented, averages across the rural–urban continuum URCA-defined categories are population weighted averages, while average across countries are simple averages, following the methodology used in Chapter 2 for the calculation of regional CoAHD.

back to top