Natural Resource Exploitation and Rent-Seeking Behavior: An Experiment with Local Authorities and Charcoal Producers in Burundi

We will conduct three lab-in-the-field experiments with participants in rural Burundi. In the first two interventions, local authorities—village chiefs and other village-level authorities—and charcoal producers will engage in a dynamic game. This game simulates a common form of bribery in rural Burundi, involving the overexploitation of trees for charcoal production. The third intervention will use the same experimental design but with farmers from the same region as the other participants. The goal of this last experiment is to compare the behavior of charcoal producers and local authorities with that of the general population.

In addition to these three main experiments, we will conduct smaller experiments and surveys with our participants. We will gather data on their demographic characteristics, their valuation of forests and the environment, and their levels of trust. We also plan to run experiments to assess their risk aversion and honesty levels. Finally, we will collect information on deforestation and reforestation in their villages of origin.

2024-06-18

2024-09-18

Key outcome variables:
- Fequency at which charcoal producers (exploiters) offer a bribe to local authorities;
- Bribe size;
- Frequency at which local authorities accept the bribe.

Our main experiment consists of a simple dynamic game with two players. Player A, the exploiter of the natural resource, can derive revenue from a local natural resource—trees in our framing—either legally or illegally. If Player A chooses the legal way, the game ends, and Player A receives their natural resource revenue minus a government tax for exploitation, plus a small initial endowment. Player B receives only an initial endowment, representing their salary.

If Player A decides to go the illegal way, they must offer a bribe to Player B. Player B can then decide to accept or reject the bribe. If Player B rejects it, Player A loses their initial endowment and does not receive any payoff from the natural resource. If Player B accepts the bribe, Player A pays the bribe and receives the natural resource revenue. However, if Player B accepts the bribe, there is a small chance that both Player A and Player B will be caught by higher authorities and lose all their payoffs.

This game is played by real local authorities (Player B) and exploiters (Player A, charcoal producers, who cut trees to make charcoal and sell it in the villages). The parameters of the game are calibrated so that the revenue from the natural resource is equivalent to 3,000 FBU, or 1 USD. This is equivalent to one afternoon of work in rural Burundi.

Pairs play the game face-to-face, with no oral communication. All interactions are conducted in writing. Players play the game several times but never with the same other player. Each experiment is run in a single session, with all participants present on the same day.

We have one control group and six treatment groups. In the first five treatment groups, we vary key parameters of the game, such as the initial endowment, probability of being caught, amount of the fine if caught, etc. The last treatment provides information on the political affiliation of the exploiter to the local authority.

Public lottery. Pairs are chosen randomly and are grouped into treatments randomly as well.

Individual

Yes

We will have seven groups (one control group and six treatment groups), with approximately 30 participants per group and 15 pairs. Each participant plays the game an average of eight times, resulting in about 120 observations per group.

Each experiment will have approximately 840 observations.

15 pairs (30 individuals) per group. We intend to have roughly the same number of participants per group.