INTRODUCTION One of the core challenges in operationalizing effective poverty reduction programming is ensuring that investments reach the intended populations. Targeting is the mechanism that is used for “identifying households or individuals who are defned as eligible for resource transfers and simultaneously screening out those who are defned as ineligible” (Sabates-Wheeler, Hurrell and Devereux, 2015). Widely used targeting methods include proxy means testing (PMT), geographical and categorical targeting, community-based targeting (CBT) and hybrid combinations of these methods (for a summary of the main methods, see Box 1). Although the motivation for selecting a particular targeting method varies, the choice will typically depend on a programme’s objectives, while being constrained by the trade-off between the accuracy and costs associated with each method. Academics and practitioners are still debating the superiority – in terms of accuracy and costs – of CBT and PMT, two of the most commonly used targeting methods in sub-Saharan Africa. The few studies that directly compare the two methods fnd that neither one clearly outperforms the other in terms of accuracy, or that PMT performs only slightly better than CBT (Alatas et al., 2012; Karlan and Thuysbaert, 2013; Pop, 2015; Stoeffer, Mills and del Ninno, 2016). Targeting based on community knowledge, however, typically results in higher levels of satisfaction and greater perceived legitimacy of the process among recipients (Alatas et al., 2012; Pop, 2015). The evidence surrounding the performance of CBT is mixed. On the one hand, studies focusing on CBT highlight the risks of ‘elite capture’, rent-seeking behaviour and patronage from local leaders; on the other hand, they point to the potential benefts associated with local agents having better knowledge of the community, the fostering of greater community ownership and satisfaction, and the fexibility of being able to customize to the local context the criteria used to identify the poor (Alatas et al., 2012; Conning and Kevane, 2002). TARGETING OF SOCIAL PROTECTION IN ETHIOPIA Targeting in Ethiopia has received substantial attention since the country is one of the largest recipients globally of donor funds for development and emergency interventions. Several studies examining the country’s early social protection programmes have revealed targeting biases in regard to demography, geography and political affliations (Broussard, Dercon and Somanathan, 2014; Jayne et al., 2002). Innocenti Research Brief 2018-23 Political Connections No Longer Determine Targeting of Social Protection: A successful case study from Ethiopia Elsa Valli, UNICEF Office of Research - Innocenti Ethiopia is one of the world’s largest recipients of donor funds for development and emergency interventons. As such, its targetng of social protecton has received substantal atenton. In partcular, concerns have been raised that politcal connectons could play a role in determining the selecton of benefciaries. With the introducton in 2005 of the Productve Safety Net Programme (PSNP), Ethiopia implemented various policies aimed at increasing transparency in the targetng of social protecton. This case study compares targetng before and during the implementaton of PSNP, and shows improvements in targetng for both public works and emergency aid in relaton to the dimensions of poverty, food security and politcal connectons. Most notably, politcal connectons are no longer found to determine the receipt of benefts during the implementaton of PSNP.