Impact of Watchdog Journalism on Public Service Provision in Tanzania
- Politicians and electoral candidates
- Rural population
- Citizen satisfaction
- Elected Official Performance
- Information
- Community monitoring
In many low-income countries, citizens have limited access to public goods and services such as healthcare, education, and infrastructure. The rapid expansion of local and independent journalism in low-income countries may offer a promising avenue for improving government accountability and service provision, but rigorous evidence about the effect of investigative journalism is scarce. In this randomized evaluation in Tanzania, researchers examined the impact of local watchdog journalism investigations on public service provision. Local journalism improved government performance and service delivery, likely by helping central governments monitor bureaucrats at the district-level.
Policy issue
In many low-income countries, citizens have limited access to public goods and services including healthcare, education, and infrastructure. In theory, investigative journalism may be a promising tool for improving government responsiveness to citizens' needs. Journalism may inform governments about service delivery problems, or it may allow citizens to elect more effective public servants and motivate politicians once they are in office.
Central governments in developing countries often have limited oversight over local decision-making. In these contexts, district-level officials frequently misuse resources and funds. At the same time, many of these states have experienced a dramatic expansion of local and independent media since 1990. In this context, local media investigations have the potential to play an important role in exposing ineffective local governance and motivating central government responsiveness. However, few studies have rigorously evaluated the effect of local watchdog journalism on government performance.
Can local watchdog journalism improve government performance and responsiveness to citizen needs outside of developed democracies? Through which avenues might journalism investigations impact government accountability?
Context of the evaluation
In Tanzania, the single political party Chama Cha Mapinduzi (CCM) has been in power since 1964. Local Regional Commissioners maintain oversight of infrastructure projects and have influence over the promotion trajectories of lower-tier bureaucrats, though their oversight is often limited. This gives local bureaucrats the opportunity to siphon funds or neglect their duties. Further, citizens often remain unaware of which government bodies are responsible for providing different government services. Consequently, access to public goods and services remains low in Tanzania: at the time of this study, only 21 percent of Tanzanians had access to piped water, and 37 percent had access to healthcare in their community.
Over the past thirty years, local and independent media in Tanzania has significantly expanded from two radio stations in 1993, to 211 in 2022. Local radio stations routinely air news reports on community grievances about local service delivery challenges such as broken water points, absent teachers, and missing medical supplies. While the large majority of local radio stations broadcast local news daily, local grievances are rarely covered in press reports—especially in rural areas. Thus, Tanzania’s media sector remains underdeveloped compared to other countries in East Africa.
To be eligible for the study, communities had to have pre-existing problems with public service delivery—such as water, infrastructure, education, and health—of which at least some community members were aware. On average, these communities had populations of 1,000 to 2,000, and were 21 kilometers from the nearest paved road and 4 kilometers from the nearest school. Nearly half of communities in the sample reported experiencing water shortages, followed by issues with healthcare (23 percent), infrastructure (11 percent), and education (9 percent).
Details of the intervention
Researchers conducted a randomized evaluation in Tanzania to examine the impact of local watchdog journalism investigations on public service provision in partnership with fifteen local radio stations from fifteen out of Tanzania’s 26 mainland regions, each identified as having high-quality news reports by local journalism professors. Researchers randomly assigned 103 pairs of communities (206 total) to either receive the Wakati Ni Sasa intervention (intervention group) or not (comparison group).
The intervention, titled Wakati Ni Sasa (The Time is Now), consisted of three components.
- First, station-based journalists investigated public service delivery problems by soliciting community input from listeners via text messages and by visiting local communities and interviewing its leaders and citizens.
- Second, the journalists produced and broadcasted 20-minute news reports about investigation results, including interviews with citizens and government officials discussing the issue, the government’s reasons for inaction, and planned next steps. All reports were broadcasted twice a week, using a standardized template that leveraged best practices of high-performing Tanzanian journalists. Journalists informed bureaucrats that the episode would be aired in advance of the broadcast, and that they would conduct follow-up investigations in the coming months to assess the community’s progress.
- Finally, three to four months after the first broadcast, researchers conducted follow-up interviews with bureaucrats to track government responses to the investigations.
To assess changes in public service delivery, auditors conducted unannounced audits in all 206 communities five to seven months following the original radio broadcast, interviewing local leaders and key informants about their experiences with the issue, the government’s responsiveness, and their expectation that future investigations will take place in their community. Then, auditors scored these responses on a scale from -10 (the issue had become much worse) to 10 (the issue was fully solved). For example, a score of 0 indicated no change; 1-3 represented positive but small improvements; 4-6 indicated more substantial improvements; and 7-10 represented the government’s progress from start to completion of the project.
The project raises several ethical considerations that researchers considered carefully before implementation. First, researchers were initially concerned about the possibility that investigative reports would threaten the safety of investigative journalists. To reduce this risk, the researcher spent more than 12 months shadowing local journalists how they select topics and design their investigations to avoid being threatened or jailed by authorities. Second, it would be unethical to deny a reporter the opportunity to report on a given community or topic if they believed it was important. As a result, the research design provided financial support for supports in treatment villages but did not prevent journalists from reporting on problems in control communities.
Results and policy lessons
Increased local media investigations led to noticeable improvements in public service provision, seven months after the original radio broadcast was aired.
Take-up and response to program: Among the 103 intervention communities, three refused to cooperate with Wakati Ni Sasa journalists and did not allow them to broadcast their investigations. Meanwhile, the intervention left a lasting impression on local political leaders in the intervention group, as 63 percent of local leaders believed that a journalist would visit their community if it had a problem in the future compared to 46 percent in control communities. This suggests that local journalism investigations were not novel or isolated to the study, although the intervention increased citizens’ perception and awareness of investigations.
Impact on public goods and services: Audits in communities that did not receive the Wakati Ni Sasa intervention showed the original problem usually remained unchanged (37 percent of communities) or had become worse (8 percent of communities), with an average audit score of 2.03 on a -10 to 10 point scale—representing only a minor improvement in the issue. Intervention communities had higher audit scores than comparison communities in 50 percent of the community pairs, and lower scores in 35 percent of pairs. On average, intervention communities received audit scores that were 0.68 points higher than comparison communities, equivalent to one additional road or water point repair in every four intervention communities.
Impact on government and citizen responsiveness: Audit scores show that the Wakati Ni Sasa intervention increased the average government response by approximately 0.5 points, with no noticeable effect on the responsiveness of citizens, civil society, village chairpersons, ministries, or Members of Parliament. Rather, the intervention primarily spurred action by the Tanzanian Rural and Urban Roads Agency, the Ministry of Rural Water Supply, and the health and education ministries, with strongest effects in the areas of roads, water, and the environment.
Overall, results show that local journalism spurred action by government ministries in response to community issues. As investigative reports led to improvements in public service delivery, and cost under US$100 per report, this suggests that investments in local investigative journalism may be a cost-effective method of strengthening government accountability and responsiveness outside of developed democracies.
Regional radio stations have used the findings from the study to roll out a new nation-wide program, Afya Yako, to use local investigative journalism to improve health service delivery in Tanzania. The project is being implemented with financial support from the New Venture Fund and is being conducted as a randomized controlled trial in collaboration with researchers and the Economic and Social Research Foundation.
Groves, Dylan. "Local Media and Government Responsiveness: Evidence from Tanzania." Working Paper, November 2022.