Lessons from the Forest Game: Cooperation under Risk and Uncertainty

Peter Bednarik, of the Evolution and Ecology and Risk, Policy and Vulnerability programs, designed a computer game based on the forestry sector which can be used to investigate the conditions under which a “tragedy of the commons” would be avoided.

Peter Bednarik

Peter Bednarik


In many situations, public resources can be accessed by individuals without, or almost without, any restriction. Such common goods include global climate, clean air, the internet, and for all living resources with shared ownership, such as stocks of fish and game. In principle, such common goods are threatened by over-exploitation by individuals. In the worst case, this may lead to the collapse of the entire resource, a “tragedy of the commons”. This term was coined by the influential work by Garret Hardin (1968), who argued that such a tragedy is inevitable in many cases. Yet, reality shows that many common resources can be managed consistently and without collapse for a significant amount of time. A possible explanation for this difference is that often models of managing the commons lack the necessary social context. Within the cross-cutting project “Equitable governance of Common Goods” we extend this research. To this end, we designed a Common Pool Resource game in the context of forestry. Participants can harvest trees to generate income. Additionally, the forest serves as a protection against floods. Our main hypothesis is that this additional feature will cause participants to be more careful, and therefore more cooperative in managing the forest.     


We conducted experiments with 200 undergraduate students from the University of Vienna in the computer lab hosted by the Vienna Center of Experimental Economics. The students were seated in front of computers separated by opaque boxes and given written instructions. Essentially, the computer game consisted of harvesting trees from a field of patches. Each round, participants had one minute to harvest as much as they liked and another minute to view results of harvesting, including possible floods. A chat box was active throughout the entire game allowing them to communicate. The main two conditions (treatments) were (a) with flood and (b) without flood. In (b), there was, in each round, a certain amount of rainfall which could result in flood damage if not enough trees were left over to mitigate its effect. Each round, a proportion of the trees could regrow. The game lasted for 20 rounds.

Figure 1.

Preliminary results

The danger of floods significantly increased cooperation in maintaining the resource. However, the resulting increase in efficiency was not sufficient to compensate for occasional losses from floods. Hence, the final payoffs where similar in treatments with and without floods. Contrary to the differences within the treatments without floods, all treatments with floods showed remarkably similar behavior, although notable differences occurred at the start and the end of the game (detailed in a forthcoming paper).


Peter Bednarik is an Austrian citizen, and a IIASA-funded Postdoctoral Scholar (February 2015-February 2017).

Print this page

Last edited: 02 March 2016


International Institute for Applied Systems Analysis (IIASA)
Schlossplatz 1, A-2361 Laxenburg, Austria
Phone: (+43 2236) 807 0 Fax:(+43 2236) 71 313