Summer 2011 Philip Keefer and Stuti Khemani

Access to community radio improves household decisions but not government accountability in villages of northern Benin

In many developing countries citizens are unable to hold governments accountable for better service delivery, but they also underestimate the benefits of the services that government offers. Donors have long believed that access to mass media addresses both problems. They finance community radio, especially in Africa, to reinforce accountability in new democracies, and they support mass media programming to encourage households to change their health practices and to send their children to school. A new study by Keefer and Khemani, using information from 4,200 households in 210 villages across northern Benin, examines for the first time whether media access affects service delivery outcomes and, if so, whether through changes in household behavior or in government accountability.

Among the sample villages, results of literacy tests given to 2,100 second-grade children were significantly better in villages with greater community radio access. But the study could identify media effects only on household behavior: households with greater media access spend more on their children’s education. In contrast to previous research and to the main rationale for donor efforts to improve mass media, the study found that community radio access did not appear to increase the ability of villages to demand greater government accountability. Government inputs into education were not significantly different across villages with more and less access to community radio.

These results are consistent with those of a companion investigation by the authors, using data from the same survey but focusing on household acquisition of mosquito bed nets.Households with greater access to community radio are significantly more likely to pay for mosquito bed nets.But they appear to have lower access to free, government-provided bed nets: in areas where household demand for bed nets is higher, more bed nets are diverted from the free program into a gray market, where households pay for them.

Is the relationship between community radio access and service delivery a spurious association, or does the first cause the second? This is a serious concern: community radio stations broadcasting about health and education may prefer to locate in areas where households are already more concerned about these issues.Unobserved levels of household concern about health and education, rather than community radio access, would then explain the association between community radio access and health and education outcomes. In fact, the conditions of the radio market in northern Benin make it clear that village access to community radio is unrelated to village characteristics.

Small community radio stations with a limited broadcast radius dot northern Benin. Even within communes, villages vary substantially in the number to which they have access.This variation is unrelated to village characteristics that might also influence service delivery. When radio stations have weak signals, insignificant topographic and geologic differences between villages can have a substantial effect on their radio access. As a result, as the sample of villages shows, a village may receive the broadcasts of fewer radio stations than an otherwise identical neighboring village. It is implausible that such differences would affect service delivery outcomes.In addition, historical accounts of the founding of community radio stations indicate that highly local factors (such as intracommune power struggles), affecting only villages in the communes where the stations are located,determined their location. The study’s conclusions, however, are based on within-commune comparisons of village access to out-of-commune community radio broadcasts.

Surely, though, one might argue that radio stations’ programming choices are related to the characteristics of the villages in their broadcast area. After all, it is well known that community radio stations cater to the linguistic and cultural demands of their listeners. But they receive little or no revenue from advertising. Instead, they depend heavily on the financing they receive from donors and government ministries in exchange for broadcasting programs that suit the priorities of the donors and ministries (health and education), not those of listeners. Villages with greater community radio access therefore have greater exposure to broadcasts touting the virtues of better health practices and the importance of education.

These results have three policy implications. First, lack of policy information is often not the only impediment to collective action needed to hold governments accountable. As in Benin, citizens may lack organizations, particularly political parties, that can mobilize even fully informed citizens for collective action. Second, community radio stations that depend on donor and government financing may be particularly unlikely to broadcast "accountability" programming. Third, even in the absence of an accountability channel, media interventions can still influence household behavior. But media outlets that can support themselves with advertising may be reluctant to broadcast the health and education programs that influence behavior in northern Benin.

Philip Keefer and Stuti Khemani. 2011. "Mass Media and Public Services: The Effects of Radio Access on Public Education in Benin." Policy Research Working Paper 5559, World Bank, Washington, DC.

____Forthcoming. "The Role of Mass Media in Developing Countries: The Impact of Radio on the Distribution of Anti-malaria Bed Nets in Benin." Policy Research Working Paper, World Bank, Washington, DC.