Brief 51: Does Information Technology Improve Public Service Delivery?
EGAP researchers: Guy Grossman, Melina Platas
Other authors: Jonathan Rodden
Geographical region: Africa
Research question: Can free and anonymous information communication technology strengthen local accountability and improve the delivery of public services?
Preparer: Catlan Reardon
The study, which is forthcoming at World Development, takes place in Uganda, a low-income East African country with a high level of decentralization. Specifically, the intervention was implemented in Arua district, a rural and fairly poor area of northwestern Uganda. In rural areas in developing countries, representatives often lack information about the quality of public services on the ground. This information shortfall is due to the high cost for the government to actively monitor service points, as well as the high cost to citizens who observe problems directly and report service delivery problems. The study aims to improve service delivery by leveraging a new information communication technology (ICT) platform that reduces the monetary and social cost of monitoring service providers and sharing information with local government officials.
The U-Bridge program, which allowed citizens to send free and anonymous text messages directly to local government officials, was implemented by Governance, Accountability, Participation and Performance (GAPP) program in conjunction with UNICEF, and funded by USAID. Unlike other ICT programs, U-Bridge primarily targeted local bureaucrats rather than local politicians. This is because in Uganda, local bureaucrats are in charge of implementing public services and managing resources as well as overseeing front-line service providers.
The authors implement a randomized field experiment to test whether a text message-based ICT program – U-Bridge – improves local government monitoring, frontline service provider efforts, and availability of inputs in education, health, and water in Arua district, Uganda. Researchers first created clusters of villages around Arua’s 48 mid-level government health centers (the unit of randomization). Next, they randomly assigned half of those clusters to the treatment group and the remaining to control. The treatment intervention involved holding quarterly community meetings to inform citizens of the new technology and door-to-door registration. The community meetings disseminated information not only about the technology itself, but also about national standards for service delivery and the roles and responsibilities of local government. With 4-5 villages in each cluster, the sample included 243 villages – 131 of which were allocated to the treatment group – and ultimately more than 3,000 registered users.
Primary outcome measures were drawn from both administrative data and unannounced audits across three domains: monitoring, effort, and inputs. The study focused on three sectors: health, education, and water. Government monitoring was measured as local government inspections and calls to clinics and schools, effort via rates of staff absenteeism, staff meetings, outreach clinics, and teacher engagement, and inputs included availability of medicines and school supplies. Additional data on the use of health services and school test scores also served as outcome measures. Service delivery in the water sector was measured by the number of water parts/services distributed and water-related requests. Importantly, study authors were interested in both short-term and longer-term effects. Thus, measures were collected after both one and two years.
At the end of the first year, the authors find positive and to the most part significant effects for the education outcomes – particularly, for monitoring and inputs. However, education improvements wane after the first year. For the water domain, the authors find evidence of a relatively large positive effect, on water parts and services distributed, but this relationship fell slightly below conventional levels of statistical significance. On the other hand, the ICT intervention did not have any appreciable or consistent effect on health service delivery at any point in time.
Interviews and focus groups with users and district officials, as well as an endline survey distributed in selected treatment villages, suggest that citizens were often disappointed with their officials’ responses to their messages, leading to a decline in program usage over time. Moreover, they find that awareness of the U-Bridge at the close of the program was not relatively high – only about one third of potential users reported ever hearing about the program. At the same time, district officials reported that while U-Bridge revealed some new information about service delivery issues, many of the messages they received did not provide specific enough information about the problem in question, or were unrelated to their purview of governance. The authors conclude that while there exists demand for such platforms, the program only resulted in piecemeal improvements. There was no general improvement in service delivery outcomes in response to the program.
Increasing access to mobile phones across the developing world suggests possible technological solutions to some of the most intractable governance challenges, ranging from teacher absenteeism to drug stockouts. The U-Bridge ICT platform implemented in rural Uganda shows that while there was high uptake of the program, its effects on service delivery outcomes were limited. The initial take-up of the program – over 11,000 messages sent – reveals that there exists a high demand for governance-related ICT programs. Yet despite the high volume of messages, the number of actionable and relevant messages to which district officials could actually respond was low, suggesting that citizens need training regarding the type of information officials need in order to solve service delivery problems. At the same time, many users of the platform were disappointed in officials’ responses, pointing to mismatched expectations for how problems would be solved in response to the program. Suggestions for similar programs in the future include involving multiple layers of local government, targeting specific service sectors, third-party monitoring of government responses and offering a set of templates for user reports instead of open-ended messages.