Experimental Design
We conducted a lab-in-the-field experiment in Liberia using tree extraction games that mimic decision making in real-world natural resource management. This experimental framework seeks to shed light on the impact of community discussions and enhanced policy comprehension regarding Liberia's Land Rights Act of 2018 on resource management, especially when faced with resource shocks. The lab-in-the-field tree harvesting game series took place between 16 July and 15 August 2018. The game featured 280 participants from 270 households, organized into 48 teams across 12 towns or villages in 10 counties. Each team, comprising 4 to 6 participants, was assembled randomly. Concurrently with the game, a comprehensive household survey was carried out from 19 May to 03 December 2018, encompassing over 1,300 individuals, including game participants. The geographical distribution of the Clans where the research took place in Liberia is on Figure \ref{fig:game_liberia}. Notice that, the spatial distribution of this research covers pretty much every part of Liberia which is a good sign of representativeness of the whole country.
All decisions were made in private to avoid influencing other participants' choices, and discussion between participants was only permitted in treatments where communication was explicitly part of the intervention. In the beginning, household IDs were recorded to enable the linkage of game behavior to household survey data. Participation in the experiment was voluntary. We mimic real-world incentives to cut down trees by providing monetary compensation equivalent to 1 Liberian dollar ($ 0.0056 USD) for each tree harvested as an immediate financial reward. Participants were also briefed on the purpose of the game and consented to participate, understanding that payment depended on completing all rounds.
Every game arm either starts with 60 trees (regular game) or 30 trees (shock game), with regrowth of trees, depending on the remaining number of trees between rounds. During every game arm, for every 10 trees left unharvested after a round, 2 new trees are added to the stock to simulate the regrowth of trees, reaching a limit of 160 trees in total. If the number of remaining trees fell below 8, participants were restricted from further harvesting, simulating ecological limitations.
We implemented seven treatments in sequence during the game. These treatments simulated various forest management scenarios, ranging from completely open access (baseline) to scenarios with restricted access and facilitated conditions, including organized information sharing. The game began with a practice arm, where participants got accustomed to the game's mechanics in an environment without any treatments, followed by an open access round, the same environment as the practice round, starting with 60 trees. The maximum number of rounds for each game is between 1 and 6. For analytical purposes, the two initial arms were merged into one baseline category. This combined game arm represented a strictly open-access scenario, serving as a benchmark for evaluating all subsequent treatments. It featured stable resource levels and minimal directives, simulating an open-access resource management setting.
Next, we implemented a resource collapse scenario, which aimed to create an abrupt environmental disturbance by reducing the initial forest size to half, 30 trees, and restricting each participant's maximum harvest to four. The importance of this treatment lies in examining how participants adjust their harvesting decisions when unexpectedly faced with a scarcity of resources, a situation that is increasingly occurring with climate change and extreme weather events. Subsequently, we analyze the impact of group communication on resource management through the "Communication" treatment. In this stage, participants were allowed to communicate in a five-minute group discussion about their harvesting strategies before making individual decisions. This intervention sought to explore how community dialogue affects sustainable resource utilization and collaborative decision making.
The game then introduced a structured policy awareness component through the "Information" and "Facilitation" treatments, centered around a poster detailing LRA reform. This policy, which serves as the foundation for land reform in Liberia, emphasizes equitable access, environmental stewardship, and community rights to customary land. Specifically, the policy delineates four types of land (public, government, private, and customary land), and advocates for equal protection of customary land, local governance, and community benefit. This was particularly relevant to the participants, as it informed them about land tenure, rights, and governance principles that closely align with their lived experiences in managing communal resources.
Information treatment involved reading the poster on LRA aloud to participants without further explanation. The participants had then five minutes to discuss their decisions among themselves, but without the opportunity to ask questions or seek clarification of the policy. This treatment aimed to observe the effect of passive exposure to land use rights information on harvesting behavior. In the facilitation game, the poster content was read and explained in detail, and the facilitator, who is an NGO personnel working on land use rights, actively led a discussion and answered any questions posed by the participants. This round was designed to gauge the impact of the facilitated information on land use rights on resource extraction decisions.
The final intervention, "Shock After" treatment, combined the elements of all previous treatments (communication, information, and facilitation games) with a second resource shock identical to the one in "Shock Before". Starting with only 30 trees and a maximum allowable harvest of four trees, this treatment simulated a resource collapse condition but with communications and facilitated information on land use rights. This setup allowed for a nuanced analysis of decision-making under compounded conditions of resource scarcity, community dialogue, and facilitated information on land use rights.