Katherine Casey, Rachel Glennerster, and Edward Miguel (2011). NBER Working Paper 17012, May.

This randomized controlled trial of a community driven development program in Sierra Leone casts doubt on this popular intervention method, finding few long-term effects on public goods provision or the inclusion of marginalized groups in politics.

Economists have long acknowledged that institutions have a central role in moderating economic performance.  For instance, it is widely acknowledged that institutions that protect property rights are correlated with better economic growth, and former European colonies are often endowed with strong institutions relative to their counterparts.  Despite this importance, economists know relatively little about the tactics with which these institutions can be reformed.  Casey, Glennerster, and Miguel write that “there is no consensus on the specific policy reforms that will successfully engender better functioning institutions, or on whether it is possible, or even desirable, for external actors like foreign aid donors to attempt to reshape local power dynamics in less developed countries.  This debate has played out vigorously in discussions on aid policy reform: while some scholars argue that large infusions of foreign air can themselves help build stronger institutions (Sachs 2005), others assert that historically rooted local institutions and social norms are difficult to understand, let alone transform (Easterly 2006), and that attempts by outsiders to create ‘better’ institutions will be futile” (1).

“Community driven development” is one of the most popular tactics for improving government institutions in the development community.  It aims to improve the accountability, competence, and inclusiveness of local institutions generally through block grants for local public projects combined with training designed to promote effective collective action.  Coordination and participation are paramount, encouraged through village development committees and the inclusion (particularly in leadership roles) of women and marginalized groups.  Community driven development generally falls under the category of bottom-up, decentralized reform.

Skeptics of community driven development suggest several critiques.  These include that “CDD participation requirements serve as a regressive tax, widening political participation will clog up rather than expedite decision-making (Olson 1982), external resources may attract new leaders and crowd out the more disadvantaged (Gugerty and Kremer 2008), and that these additional resources will be captured by elites if the program is unable to change the nature of de facto political power (Bardhan 2002)” (2).

Despite compelling arguments for and against community driven development, there have been few empirical evaluations of its effectiveness.  In this paper, Casey, Glennerster, and Miguel use a randomized controlled trial to evaluate the impact of a community driven development project in Sierra Leone.  Sierra Leone’s recent civil war left the country at the bottom of the UNDP Human Development Index.  In an effort to rebuild, the government launched the World Bank-funded GoBifo (“Move Forward”) program in 19 district councils.  The program provided block grants of $4,667 to communities, averaging $100 per household, to provide public goods, seed capital for businesses, and training in trade skills.  26% of village projects involved agricultural initiatives such as seed multiplication and communal farming, 17% skills training and small business development initiatives in blacksmithing, carpentry, and soap making, 14% livestock or fishing, 14% the construction of community centers or sports fields, 12% primary school repairs, 10% water and sanitation, 5% health, and 2% roads.  The authors report that “GoBifo’s design is similar to many other CDD projects in non-post-conflict societies” (3).

The authors conducted a randomized evaluation of 236 villages comprising 2,832 households in the program between 2005 and 2009.  Specifically, they assessed the degree to which treatment versus control communities could successfully deal with “structured community activities” after the GoBifo program had ended—taking advantage of an opportunity to buy subsidized building materials, making a communal decision about the purchase, and allocating the materials within the community.  This is a noteworthy methodological innovation and the authors report that they “feel that these SCAs capture actual local collective action capacity, and uncover the decision-making processes that underlie it, more accurately than lab experiments, hypothetical vignettes or survey reports alone” (4). 

The GoBifo program was well managed: village organizations were established, financing was disbursed, and leakage of funds was minimal.  Treatment villages exhibited more and higher quality public goods.  However, the authors report that there is “no evidence that the program led to fundamental changes in the ‘software’ of collective action—namely, local fundraising capacity, decision-making processes, or even social attitudes and norms” (5).  While women were included in GoBifo projects, their post-program likelihood of voicing opinions in community meetings or acting as leaders had not increased.  Treatment villages were not better at fundraising than control villages. 

These results depart from some earlier findings, such as a study reviewed previously in this series by Lori Beaman, Raghabendra Chattopadhyay, Esther Duflo, Rohini Pande, and Petia Toplova in 2009 that found that quotas for female leaders in Indian village councils decreased statistical discrimination against female candidates in only a few years.  These results are consonant, however, with varied results in recent impact evaluations of CDD programs.

These findings represent an important check on the optimism surrounding community driven development.  In the short term, local public goods provision increased, but there were no sustained impacts on fund-raising, local decision making processes, or the involvement of marginalized groups such as women in local affairs.  The authors conclude that the program was “ineffective at durable reshaping local institutions.”  Overall, “The results contradict the currently popular notion in foreign aid circles that CDD is an effective method to initiate social change or fundamentally alter local decision-making processes.  …While ‘good’ institutions may be critical for successful economic development, our findings provide another piece of evidence that institutions and social norms are difficult to change” (32).