Deriving Sub-national Poverty Estimates for Sub-Saharan Africa

HarvestChoice in collaboration with the International Center for Tropical Agriculture (CIAT) produced a new series of sub-national poverty maps for sub-Saharan Africa (showing the prevalence of individuals living on less than $1.25/day and less than $2/day) . These new maps supercede HarvestChoice's global map of poverty first published in August 2010.

Method and Sources

In this new release we derive all sub-national poverty distributions from nationally-representative household surveys conducted around the year 2005. Table 1 below provides the source and year of all country micro-data used (29 countries as of SChEF r12.04). The poverty ($2/day) and extreme poverty ($1.25/day) lines are expressed in international equivalent purchasing power parity (PPP) dollars.

In all surveys we reproduce the sub-national distributions of per-capita consumption expenditure (a synthetic indicator expressing the money-metric welfare utility level) and then shift the national poverty line up or down to coincide with the published PPP $1.25 an $2/day international lines for that particular survey (Povcalnet). This results in adjusted poverty prevalence rates at the sub-national level (typically at administrative level-1 or lower depending on a survey's representativeness).

As such, poverty rates refer to the actual survey year, although they are expressed at the same internationally comparable 2005 PPP level. Since estimates are based on local currency valued at 2005 PPP, poverty ratios are therefore comparable across countries.

The method is illustrated in Figure 1 below.

Figure 1: Regional distributions of per-capita consumption expenditure and international poverty lines. Mozambique, 2002-03. Click to enlarge.

Gridded Maps of Poverty and Extreme Poverty for sub-Saharan Africa

Sub-national poverty estimates are further rasterized to HarvestChoice 5 arc-minute grid. Poverty rates are assumed uniform across all gridcells belonging to the same sub-national unit. Poverty rates in countries for which no survey micro-data is yet available are set at the 2005 national level reported by PovcalNet (World Bank, 2012).

Table 1. Countries with national household surveys and censuses used in gridded poverty headcount ratio calculations.

Note that poverty headcount ratios were estimated at level-1 administrative units (regions or provinces) in all surveys apart for KEN-IHBS-05 for which estimations are possible at level-2  (Wilaya) due to a higher sampling size.

Maps 1 to 3 below show the reference year of each national survey (Map 1), the '05 PPP $1.25 and $2 poverty headcount ratios for all survey administrative units (Map 2), and their corresponding standard deviations (Map 3).

Reference years of source national surveys (credits: Elizabeth Barona, CIAT)

Map 1: reference years of source national surveys (credits: Elizabeth Barona, CIAT).

Poverty HCR at $1.25 and $2 and corresponding standard deviations (credits: Elizabeth Barona, CIAT).

Map 2: poverty HCR at $1.25 and $2 and corresponding standard deviations (credits: Elizabeth Barona, CIAT).

Poverty gaps at $1.25 and $2 (credits: Elizabeth Barona, CIAT)

Map 3: poverty gaps at $1.25 and $2 (credits: Elizabeth Barona, CIAT).

 

Further refinements are under way to line-up all country estimates at the same point in time (not all current data points refer to 2005, with a maximum time lag of +-2 years for a limited number of countries).

 

 

Citation

HarvestChoice, 2012. "Deriving Sub-national Poverty Estimates for Sub-Saharan Africa." International Food Policy Research Institute, Washington, DC., and University of Minnesota, St. Paul, MN. Available online at http://harvestchoice.org/node/1932.

Jan 17, 2012