On the mechanics of the political resource curse: Behavioural measurements of information and local elite behaviour in Mozambique

Mozambique discovered substantial natural resources in recent years. Known gas reserves in the Rovuma basin have the potential to transform Mozambique into a global player in Liquefied Natural Gas exports. Being a recent democracy, and with relatively weak institutions, Mozambique also faces considerable risks of resource and revenue mismanagement in the future, particularly since media independence and penetration are low and the level of political accountability is not improving.

The intervention we propose to evaluate is the distribution of information about the management of the newly discovered natural gas in the Rovuma basin, and the holding of citizen meetings to deliberate on the main priorities for spending the revenues from these resources. These activities will be sponsored and implemented by a consortium of NGOs. These implementing agencies strongly believe that the resource curse can be counteracted through the widespread provision of information about the management of natural resources in the country. Behind this belief is the hypothesis that information will make politicians accountable through the electoral system.

Our knowledge about what works in preventing the natural resource curse in democracies is limited. There is no causal evidence about the effects of the provision of information in these contexts. In fact there are fundamental doubts about the theoretical mechanisms that link natural resources to decreased income: limited political accountability is not the only possible explanation, as there could be a generalized increase in rent-seeking activities in the economy. This impact evaluation will seek to advance on both dimensions, establishing causality and identifying the mechanisms of change. Both will enable better policy design.

Our main evaluation questions are:

  1. What is the effect of information about natural resources on the political elite’s behaviour, when local elites are the only ones being informed, when everyone is informed, and when citizens deliberate on ways to spend resource revenues?
  2. What is the effect of information about natural resources on citizens’ behaviour, when everyone is informed and when citizens deliberate on ways to spend resource revenues?
  3. What is the impact of different types of information channelled to citizens, namely on aspirations and other citizen outcomes?
  4. What is the impact of different types of deliberation, i.e., default, majority elections (secret), and open deliberation, on elite behaviour/deliberation outcomes?
  5. How are the above-referred impacts different for subjects living closer to the centre of action for the natural gas boom?

We propose to conduct a two-layer randomized evaluation. We will evaluate three different treatments at the community-level. 200 communities will be covered by our experiment. The first treatment is the dissemination of information about natural resources and its management to the local elite only. The second is the dissemination of information to both the population through community meetings and door-to-door contact, and to the local elite. The third is information plus deliberation in small citizen committees about priorities for the spending of natural resource revenues. We will also evaluate different types of information to citizens at the level of the individual (aspirations module) and different methods of deliberation at the level of the citizen group (deliberation module).


  • Research in progress.

    Project last updated on: 17 Feb 2016.