Brief 24: Reducing Elite Capture in the Solomon Islands
EGAP Researcher: AidData
Partners: Andrew Beath, Giovanna d’Adda, Pauline Grosjean, Roberto Weber, Ariel Ben Yishay
Can decentralizing local public spending decisions to the individuals who will benefit from these decisions improve the way in which funds are spent?
Preparer: Jasper Cooper
It is generally acknowledged that decisions about how to spend public money are best made at the local level where the funds will be used, as local communities best understand their own needs and can monitor whether the funds are well spent. However, how can we ensure that powerful local actors do not manipulate spending decisions to benefit their own interests, to the detriment of the community? This study tested a novel mechanism for decentralizing public funds to the local level through a randomly assigned intervention that took place in 80 villages throughout the Solomon Islands. Each village receives a grant of 2000 Solomon Islands Dollars (approximately $250 USD) to implement a community building restoration project, which is decided upon by ten females and ten males from the community. In the control villages, this grant is given as a single block, and in the villages assigned to treatment, it is given to participants as vouchers that can be traded in for private cash at half their value. The treatment increased the breadth and vigor of community participation in the decision-making process around the use of the funds, and caused more overlap between the preferences of the community at large with the project chosen by the group. The treatment greatly increased participation of individuals who do not usually participate, and greatly increased their perception of the fairness of the project finally chosen.
In many developing countries, decisions about how to spend public money are made by the central state, but recent years have witnessed a move to decentralize spending to the local level, as local communities are thought to have a better understanding of how the money can best be spent and to be better positioned to monitor its use. However, there is concern that powerful members of the local community can nullify the benefits of decentralization because they are able to bend spending decisions toward their private interests. This study tested an innovative decentralized funding mechanism that sought to limit the extent of this phenomenon, often referred to as ‘elite capture’. From 2008 – 2013, in partnership with AusAID, IFAD, and the World Bank, the Solomon Islands Ministry of Development implemented the US$ 22 Million Solomon Islands Rural Development Program, which financed investments that villagers themselves played a large role in determining. Local institutions such as tribal councils and churches planned and managed most of the project implementation funded by this project. The villages featured in this study all benefited from the Rural Development Program. Most are small, with an average population just below 500 people, and geographically isolated from the capital. Many are accessible only by paddle-boat or outboard canoe.
The intervention was carried out during June to August of 2013, in 80 villages selected randomly in equal proportions from the Islands’ four main provinces. In each village, nine women and nine men were randomly selected to act as participants in the implementation of a community project. The eighteen were joined by one male and one female perceived by the community as holding a high social position, designated as ‘leaders’. In both treatment conditions, the twenty participants received a SI$ 2000 grant that they could spend at a local hardware store in order to upgrade some community infrastructure of their choosing (a public building, well, or irrigation system). In the control, this grant was delivered as a lump sum, earmarked for the project. In communities assigned to treatment, the 20 participants were each handed ten SI$ 10 vouchers (approx. US$ 1.4), which could be anonymously contributed to the collective fund, or anonymously traded in for half the value (SI$ 5) in cash, to be used privately. In both treatment arms, the participants came together for a designated meeting, attended by a survey enumerator, to vote anonymously on what project to implement. Before this meeting took place, participants responded to a short questionnaire about what sorts of potential community projects they preferred, and were asked to rank the three most needed projects for the community. Enumerators also collected information on the dynamics of the discussion during the meeting. Finally, a second questionnaire was administered to the participants after the meeting had taken place, in order to gauge their perceptions of the fairness of the decision-making procedure, and their satisfaction with the projects selected.
Compared to the control, in villages where individuals received vouchers meetings lasted for about three minutes longer on average, featured somewhat broader participation, and individuals spoke more frequently in the debate over how to spend the funds. The effect of the treatment on individual participation in meetings was particularly strong for those individuals that did not have a history of attending community meetings: among this group, receiving the vouchers increased the probability of participating by around 9-10 percentage points. This suggests that individual transfers may encourage and empower isolated members of a community to participate more fully in collective decision-making processes.
Moreover, the voucher system caused changes in how well the projects reflected the preferences of the community. Projects implemented in treatment communities were 5.5 percentage points more likely to correspond to the type of project that members of the community had stated a preference for in surveys prior to the intervention. Tellingly, projects implemented in treatment communities were less likely to reflect the stated preferences of community leaders, indicating that individual transfers might mitigate misalignments between the preferences of leaders and those of the community.
While individuals in the study were very likely to perceive the process of selecting and implementing projects as fair, the treatment nevertheless increased these perceptions relative to the control. Community members in treated villages were almost 6 percentage points more likely to perceive the projects as fair and satisfactory than those in the control. Importantly, treatment effects were strongest among those who do not participate regularly in community events: while such persons were initially much less likely than the rest of the community to view projects as fair, when treated their perceptions of fairness increased to a level equivalent to the average community member.
Finally, the researchers found that the policy of distribution to individuals was not without its own costs. Private cash conversions represented just over a third of all the public funds provided to the treatment group. It is difficult to determine to what extent this constitutes an inefficiency: the funds used for private consumption would have been doubled in value if used towards a public good, and in this sense, the community benefited less than it may otherwise have in the absence of private consumption. However, using one’s own funds toward a project might create a sense of ownership among individuals, leading them to maintain persistent monitoring of the project and its implementation, which might counteract the costliness of funds leaked for private consumption.
This study tested an innovative new method for radically decentralizing spending decisions at the community level. By empowering individuals with the ability to anonymously withdraw funds, this funding mechanism may limit elite domination of public goods provision in local settings, because leaders rely on cooperation and consensus in order to mobilize the available funds. The voucher treatment caused a greater degree of alignment between the preferences of citizens and the project implemented by the community, and a lesser degree of alignment with the preferences of the leader. Therefore, in contexts where elite capture is a serious threat and leads to sub-optimal allocations of resources, voucher-based decentralizations may constitute an effective anti-capture policy.
Such policies may also be desirable in communities where elite capture is not a serious concern, but exclusion of isolated or marginalized individuals is. The voucher treatment caused much greater levels of involvement among such individuals, and increased the likelihood that they were satisfied with the community decision-making process and perceived it as fair. Coupled with the finding that participants talked slightly longer and engaged more frequently in dialogue in the voucher treatment condition, this study suggests that voucher-based distributive programs, when coupled with the right decision-making framework, could lead to a greater degree of participation in community decisions.
However, policymakers should carefully weigh the benefits of such policies against their costs. In this study, most participants in the voucher condition contributed only half of their possible contribution to the public good, resulting in an under-utilization of the funds potentially available to the community. While these costs may be offset by the benefits mentioned above, in situations where local leaders can be trusted to make fair, well-informed and accountable decisions about spending on public goods, block grants may lead to more optimal allocations of the available public funds.