The Formation of Job Referral Networks: Experimental Evidence from Urban Ethiopia
36 Pages Posted: 16 Aug 2013
Date Written: August 2013
Unemployment is currently a major policy issue in many African countries (African Development Bank 2012). As pockets of widespread unemployment often persist despite labor market expansion, there is a growing demand among policy makers for research and policy tools that address poor labor market attachment. In this study we focus on exclusion from job contact networks, which constitutes a major disadvantage for labor market participants in settings where referral hiring is common and information about jobs hard to obtain. In a mid-size town in northern Ethiopia, where these mechanisms are at work, we observe that many individuals do not access local job contact networks. Models of strategic network formation and behavioral decision theory suggest that given the right incentives, job contact networks should be more inclusive. On these grounds we hypothesize that workers would link to peripheral peers when this maximizes their chances of referral and when self-regarding concerns are absent due to social preferences. We devise an experimental design adapted from Beaman and Magruder (2012) to test these hypotheses. Our results lend broad support to the assumption of strategic network formation. In a setting where competition for referrals makes connections with central players undesirable, agents tie preferably to peripheral peers. However, in treatments in which other-regarding considerations are made salient, players do not seem to choose links with peripheral peers more often than at random. Our findings suggest that a modification of field incentives, for example, via a reform in referral hiring procedures, can generate more inclusive job contact networks.
Keywords: Ethiopia, East Africa, Africa South of Sahara, Africa, social network, Labor market, field experiment
Suggested Citation: Suggested Citation