Targeting in Kenya’s Cash Transfer Programme for Orphans and Vulnerable Children

Wei Ha, Jingqing Chai and Carlos Alviar UNICEF

Understanding and measurement of poverty have traditionally relied on purely economic approaches, with income or consumption as the only indicator. Substantial progress has been made over the past ten years in the conceptualisation of multidimensional poverty and in developing methodologies for its measurement. The paper simulates the targeting performance of the Multidimensional Deprivation Index and Proxy Means Test approach using Kenya Integrated Household Budget Survey and the institutional setup of the Cash Transfer for OVC in Kenya. It reaffirms the earlier finding in the literature that there is no one preferred method or system when it comes to targeting eligible households. When we simulate the targeting performance of the MDI and PMT used previously and currently in the CT-OVC programme in Kenya, it seems that PMT consistently outperforms MDI with smaller inclusion error but at the same time, has higher exclusion error. However, PMT approach has a distinct advantage over MDI when we use the bottom 10% counting approach. It leads us to conclude that MDI can be used as a rather good geographical targeting tool when multi-indicator survey or census is available and up-to-date but not the income/household budget surveys.