Bangladesh is a country with thousands of development organizations, each with hundreds of concurrent projects. For some perspective, in 2003 there were 6,559 development NGOs operating in the country, almost one for every village. The proliferation of aid actors in every conceivable sector--from social welfare to governance and the environment--raises a fundamental question: What is actually happening on the ground? For all these projects, many of which share similar goals and locations, there is limited available information on how the aid actually assists impoverished peoples.
Last year, I worked for the World Food Programme in Bangladesh, and observed firsthand how a lack of information sharing on-the-ground can hinder effective targeting of aid projects. Since development organizations have limited resources, they generally attempt to target their projects to areas of greatest need. I assisted with the initial implementation of a flood Emergency Response Operation in Satkhira, Bangladesh, a program targeting individuals displaced by the flood as well as pregnant women and young children.
One of our tasks was to cross-check the beneficiaries selected by our NGO partner to ensure they were being chosen according to the proper needs-based criteria. As we drove along broken roads lined with makeshift shelters constructed from bamboo, jute, cloth, and plastic tarps, I anticipated meeting people in extremely dire conditions. But when we arrived at the first household on our list, a fully intact bamboo and mud house, I had a feeling that something was not quite right. As it turned out, the family had been economically hurt by the flood, but they had not been displaced. Other villagers who we talked to mentioned the same phenomenon: that certain people always seem to get selected for aid projects while others who are worse off do not. After checking a total of ten households, we found that four of them had been selected incorrectly, and reported to the WFP office immediately.
As much as development organizations try to closely monitor project implementation, it is impossible to account for every detail of on-the-ground activity. NGOs sometimes find it advantageous to selecttheir existing beneficiaries, or those who are not severely affected, in order to report significant improvements to their donors. Therefore, even if an NGO does not completely follow the implementation guidelines by targeting the intended beneficiaries, donors will have the false satisfaction of believing their project had its intended outcome. In order to identify projects where the same well-off households are repeatedly included on the beneficiary list, villagers need to be able to report on the aid they are receiving and who is being left out. Such a system would impose a check on NGO activity, pressuring them to target households with greatest need.
Increased transparency can also encourage better coordination between all of the different agencies working in a single location. In the case of the Satkhira flooding, numerous aid organizations had a presence, each with their own objectives. Everywhere I went I saw an assortment of development organization logos on shelters, latrines, and food rations. It was difficult to discern if there was any overarching coordination strategy. Indeed, as we talked to more people, they explained that some groups came in for a few days to give away bags of rice, while others had longer-term plans for building up the infrastructure and embankments. Wouldn’t it be more useful if the recipients of the development assistance could report on what aid they were receiving and what aid they still needed? Then, other donors could get a better picture of the situation and allocate funding for future projects accordingly.
In addition to improving aid coordination, transparency can enable development organizations to tailor projects to the specific needs of communities. For example, I am currently working with Innovations for Poverty Action on a randomized control trial (RCT) that is implementing demand and supply side treatments to bolster use of latrines and sanitation practices in Bangladeshi villages. As I monitored the baseline survey, I discussed the current sanitation situation with enumerators. Surprisingly, they discovered that not only was there a shortage of latrines in village households, but many village schools—built by a local NGO—also lacked latrine access. This could be for a variety of reasons. Perhaps the NGO had to build a certain number of schools and the budget did not allow for latrine construction, or perhaps they used the extra money for another project. But if the enumerators and villagers had a platform to report their observations on their latrines, donors could focus on funding the crucial sanitation component for the existing and future schools.
Crowdsourcing may be one way to address these sorts of coordination and targeting issues. Through the rapid diffusion of mobile technologies, people living and working in developing communities now have a mechanism to deliver real-time information on local conditions and project performance to donors. However, a common platform to aggregate, share, and make sense of monitoring and evaluation data does not yet exist.
AidData has recently overseen an RCT in Uganda to help develop a workable crowdsourcing model, and I hope this work expands to other areas soon. My experiences in Bangladesh have given me a new appreciation for the importance of repairing the broken feedback loop gap between donors and their intended beneficiaries. Identifying whether and how projects are functioning in impoverished communities is the central to designing effective poverty alleviation projects.