Description:

This paper explores failure in what is arguably the world’s most broadly employed mechanism design for delivering aid to poor communities in the developing world.  The World Bank estimates that 160 billion dollars has been spent over the past decade through the decentralized mechanism commonly known as community driven development (CDD).  In this study we use a case study approach to understand how CDD really works in the field; we follow the money, access to information, and incentives of participants in a number of micro-projects in Kenya that sit within a large and long-standing World Bank project. 

CDD design promises superior aid delivery in large part by solving agency problems through enhanced empowerment of beneficiaries: 1) project appropriateness is said to be improved by granting communities the right to choose their own projects; 2) project efficiency gains are predicted by letting villagers implement their own projects; 3) corruption is expected to reduce by allowing villagers to choose their leaders and use their enhanced access to information to effectively monitor all aspects of the implementation process.  Finally, the World Bank has argued that CDD benefits extend to improved upstream collective action because the social capital and empowerment built at the village level create demand for better accountability in governance. 

The micro projects examined here point to specific misaligned incentives, and reveal a pervasive pattern of corruption and failure to deliver positive development outcomes.  By creating a nearly monopolistic aid distribution structure at the national and district levels, and by setting up a repeat game of project delivery, the officers in this World Bank project hold enormous leverage over village recipients.  The project managers are effectively able to dictate terms to villagers and tightly control most information (rendering monitoring by villagers all but impossible) and whistle blowing imprudent.  In fact, the conditions created for rent seeking and project failure are considerable, and call into question the foundations of the worldwide CDD experiment.  Further, these findings have implications for the increasing devolution of development monies to local governments in Africa.