Participatory versus executive decision-making, the randomized control trial

A new paper on Sierra Leone, Chief for a Day, presented by Maarten Voors at the ABCA conference:

in a random subsample of villages the traditional elite, including the chief, were made responsible for project management. In other villages, responsibility for project management was delegated to a committee of randomly selected villagers.

While our simple theoretical model proposes that more powerful chiefs will be more corrupt by diverting more resources from the public good to their personal benefit, will under-invest in management of the project, and will seek to undermine the performance of managing committees, our empirical findings provide a more nuanced and mixed picture of the quality of local management in Africa.

Our field experiment finds little evidence that local elites managing an aid project divert more resources than the average villager, or that more powerful chiefs divert more than less powerful ones. Moreover, the village elites are able to manage a development project better than a committee of randomly selected villagers…

Projects managed by village elites are also more likely to start and be completed on time, are better constructed and maintained, and provide more (perceived) benefits for the villagers.

I think this project is more an indictment of the often naive approach to local politics in most NGO and other development programs, rather than an indictment of participatory decision-making per se.

One example: if we look at our own communities or organizations, clearly a leader or committee can be more effective if there are established channels or organizations that provide information or implement decisions, if people understand the roles and responsibilities of the decision-makers, or if those decision-makers have legitimacy in the eyes of the governed.

What that means is that changing the identity of the decision-maker may be less important if the organizations and processes and popular understanding are not changed as well. There’s a difference between changing the identity of who is in the CEO position and changing whether or not power goes to a CEO or a brand new committee.

Another example: what matters for executive decision-making is probably how constrained and accountable is the chief. Businesses prefer CEOs to committees, but that CEO is still accountable to a board, shareholders, labor law, the market, and a set of rules governing transitions of power. SO CEOs have incentives to be effective. So “chief” does not equal “despot”.

In sum, local politics may be difficult to change if it is not understood.