Institutions, peace-building and aid effectiveness in post-conflict West Africa
My current research focuses on post-conflict interventions in Liberia and Sierra Leone. Localized grievances, frustration with state failure, and lack of avenues of redress are commonly cited as root causes of the decade-long civil war in the region. The conflict left both states weakened and politically fractured, and eroded the power and legitimacy of informal agents. Strengthening inclusive institutions has become a key objective of post-conflict development strategy, and thus governments, donors and NGOs have devoted considerable resourcesto (1) improving conflict resolution and access to justice, (2) strengthening local institutions, and (3) supporting the states’ provision of public services. I am currently conducting impact evaluations of development initiatives that relate to each of these areas.
In the areas of conflict resolution and access to justice, I am evaluating the impact of a community reconciliation program in Sierra Leone that re-introduces customary reconciliation practices to war-torn communities, assisting them to set aside old grievances, build ‘social capital’, and work together for development. I am also evaluating the impact of access to justice programs in Liberia and Sierra Leone that involve paralegals who mediate disputes, advocate for their clients, and monitor the activities of customary and formal justice actors to ensure that due process is observed in the criminal justice system. In the area of strengthening public services, I am evaluating the impact of two government programs in Sierra Leone’s health sector that seek to link user feedback and monitoring to service delivery by instituting community-led ‘bottom-up’ non-financial incentives for service providers.
Customary practices for community reconciliation
Reconciliation has never been directly addressed in the development economics literature. Jointly with Jacobus Cilliers (Oxford) and Oeindrila Dube (NYU), I am conducting a randomized controlled trial of a community reconciliation program to determine its impact on conflict, social capital, and development outcomes. The intervention, implemented by a local NGO called Fambul Tok (“family talk” in Krio), aims to rebuild pre-war social institutions weakened by violence and displacement, and bring together war-torn, socially fractured communities through traditional reconciliation rituals and communal activities. The study launched in February 2011 and will continue into 2013.
The main focus of the study is to measure the impact of this intervention on a range of outcomes related to social capital. The innovative features of the design lie in the combination of standard household surveys with psycho-social measures and behavioral experiments. The behavioral experiments are designed to disentangle four different components of intra-group behavior: in-group altruism, out-group altruism, reputation and trust. Psycho-social scales measure a range of psychological impacts on outcomes deemed important for forward-looking behavior, including anxiety, depression and post-traumatic stress. Together these provide complementary assessments of attitudes, behaviors and contemporary experiences of conflict, altruism, trust and social cohesion. In addition to the reduced form impacts, a separate study will combine baseline data with geospatial analysis to identify the impact of violence on community social capital.
Community-based paralegals in Liberia and Sierra Leone
Jointly with Justin Sandefur (Center for Global Development), I am exploring the effectiveness of two new, relatively untested interventions to strengthen post-conflict legal institutions. At the core of both interventions is the provision of pro-bono legal and informational services to rural villagers with limited access to formal justice systems by paralegals trained to work at the intersection of customary and formal law. This is an alternative approach to peace-building that stresses information provision and advocacy, informal mediation and collaboration with customary authorities, and using low-skilled-labor intensive approaches to expanding legal services in remote rural areas.
In Liberia, the intervention is a randomized controlled trial of mobile paralegals who provide advice, mediation and advocacy services to the rural poor. Paralegals assist individuals and communities in negotiating the formal system. They also intermediate between the community and outside agents, be it state officials, mining companies or development agencies. The intervention has been randomly assigned across 164 villages in four of Liberia’s fifteen counties. A baseline study in 2008-09 surveyed 2,500 individuals, collecting information on conflict incidence, forum choice, outcomes, beliefs and social welfare. Communities were randomized into treatment and control groups, and treatment communities were provided paralegal visits for over two years. A second layer of randomization was then introduced at the individual level, so impact could be unbundled into (1) the paralegals’ direct impact on individual cases and (2) community-level spillovers (e.g. through community level externalities or changes in the behavior of justice providers), allowing some insight into the causal mechanisms of impact. 
In Sierra Leone, the intervention involves police and prison monitoring by paralegals who provide free legal advice and assistance to the accused. The primary focus of the project is to measure the impact of monitoring and legal aid on excessive detention, official behavior, and individual welfare. The project has also spawned two related studies that seek to understand the market for corruption in the formal justice system. The first focuses on bribe price-setting by developing a three-way bargaining model between complainants, accused and the police, using self-reported bribe payment data collected at police stations during the course of the project. The second models the supply and demand for corruption in the motorcycle-taxi market, using daily variation in rainfall as an exogenous shock to driver income (and hence the supply of bribes) to trace out the demand curve for corruption.
Community-led non-financial incentives to improve public health in Sierra Leone
Jointly with Oeindrila Dube (NYU), Johannes Haufhoser (JPAL) and Vivek Maru (Namati), I am conducting a randomized controlled trial of two interventions that aim to improve service delivery incentives in the health sector in Sierra Leone: (1) yardstick competition among clinics for non-financial awards based on utilization and user feedback, and (2) monitoring of health providers by communities using health scorecards and community-clinic interface meetings.
Most programs geared toward supply-side interventions focus on financial incentives; yet recent evidence suggests that peer recognition and status-based incentives can be more motivational, less expensive and less likely to erode intrinsic motivation. The first intervention combines yardstick competition with non-financial awards. Maternal and child health clinics will compete on service provision criteria such as vaccination rates, utilization, and user satisfaction. The best performing clinics will receive public recognition and awarded certificates, and user participation and feedback will be central to the assigning of awards.
In the second, the community takes a more active role in holding health providers accountable and working in partnership with them to improve health outcomes. Recent studies have found that community monitoring can have substantial effects on service delivery in some contexts, but not in others, suggesting that mechanism design and local context matter. In this intervention information about the state of healthcare is disseminated via a community scorecard and an action plan is developed jointly between clinic staff and the community to improve eight key indicators, including staff absenteeism, maternal mortality and vaccination rates. The initiative builds on a recent successful intervention in rural Uganda by Martina Bjorkman and Jakob Svensson in 2009. We innovate by (1) systematizing the monitoring mechanism through regularized data collection and scorecard generation; and (2) implementing at scale in partnership with the national government. The experiment will shed light on the external validity of this type of program and its applicability in other contexts.
A baseline study of 12,000 households and 250 clinics was conducted in August 2011. Clinics will now be matched along a vector of outcomes and then randomized into one of three groups: those competing for non-financial awards, those participating in community monitoring and a control group. The interventions will be launched in January 2012 across four rural districts of Sierra Leone, and implemented by the IRC, Concern, and Plan International. The endline study will go into the field in early 2013, and analysis of the results will follow. The study will assess the relative impacts of the two programs. A side study will use baseline data to develop a structural microeconomic model of health-seeking behavior by clinic users, in order to lay the foundation for the evaluation.
