Service delivery implications of census undercount in sub-Sahara Africa

Jeremy Gumbo, University of the Witwatersrand
Clifford O. Odimegwu, University of the Witwatersrand

Studies carried out in United States of America have confirmed that census undercounting compromises fair distribution of resources. None of such studies have been done in Sub Sahara Africa where most countries use census data to achieve the same goal. This study estimated effect of undercounting on service delivery in the region, using 10% samples' data from South Africa's last three censuses. The data contained adjusted counts and these were used for recreating unadjusted counts. We applied the PRINCEWATERHOUSECOOPERS method to estimate undercounting effect on the allocation of the Basic Service Grant among the country’s nine provinces using both adjusted and unadjusted counts. We further estimated the effects of undercounting on parliamentary seat allocations using the same procedure. We found undercounting to unfairly benefit certain provinces at the expense of others.