The Effects of Communication on the Partnership Solution to the Commons

Neil J. Buckley, Stuart Mestelman, R. Andrew Muller, Stephan Schott, Jingjing Zhang
2017 Environmental and Resource Economics  
Organizing individual appropriators into output sharing groups has been found to effectively solve the tragedy of the commons problem. We experimentally investigate the robustness of this solution by introducing different channels of communication that naturally arise from group competitions. In the absence of communication, we confirm that output sharing can introduce sufficient free riding to offset over-harvesting and results in full efficiency. Allowing local communication within
more » ... ing groups substantially decreases this efficiency enhancement because it reduces free riding and boosts betweengroup competition. Yet the efficiency level is still significantly higher than that achieved when global communication is allowed among all appropriators in a conventional common pool resource without output sharing. The efficiency-reducing effect of local communication is mitigated when random partners instead of fixed partners are sharing output over time, and is nearly eliminated when random partners are formed with users who belong to different communication groups. "There is more than a verbal tie between the words common, community, and communication". -John Dewey We find that in our environment, organizing subjects into groups of four players who share output introduces sufficient free-riding incentives to achieve full efficiency. Local communication within output-sharing groups decreases efficiency by countervailing the free-riding incentives induced by output sharing and enhancing between-group competition. The reduced efficiency level with local communication and output sharing is, however, still significantly higher than simply allowing global communication among all resource users without output sharing. The efficiency-reducing effect of local communication is mitigated when random partners instead of fixed partners are sharing output over time, and is nearly eliminated when random partners are formed with users who belong to different communication groups. EXPERIMENTAL DESIGN AND PREDICTIONS In each decision period every participant is required to allocate a fixed endowment of effort, e = 28, between a private activity, which provides a known return per unit of effort, r = 3.25, and a resource random) does not affect the effectiveness of the output-sharing mechanism in the absence of communication. Global Communication with and without Output-Sharing Groups This sub-section concerns the performance of the participants in the basic CPR with Global Communication, the Fixed-Partners with Global Communication and the Random-Partners with Global Communication treatments relative to the theoretical predictions for these environments. The parametric tests are based on the results reported in Table 3; for example, the effect of introducing global communication on system effort in the Fixed-Partners treatment is the sum of the coefficients of G (the global communication factor) and G & F (the interaction between the global communication and fixedpartners treatments) in Table 3. The significance of this sum, approximately 15, can be tested directly using these coefficients and their standard errors). Adding global communication to the basic CPR treatment reduces mean system effort by 22 units (about 8%) to 262.70. This effect is statistically significant in the direction predicted (t-test, p = 0.023; Mann-Whitney U-test, N1 = 4, N2 = 4, p = 0.057; both 1-sided) but a much larger reduction in effort (128 units) is required to reach the system optimum of 156. Adding global communication to the Fixed-Partners treatment actually raises mean effort by 15 units to 165 (about 10%). The effect is not statistically significant (F-test, p = 0.151; Mann-Whitney U-test, N1 = 4, N2 = 4, p = 0.20). When global communication is added to the Random-Partners treatment, mean effort rises by 16 units to 169 (again, about 10%). This effect is not significant (F-test, p = 0.222; Mann-Whitney U-test, N1 = 4, N2 = 4, p = 0.343). These relationships are displayed in Table 2 and Figure 1 (the three bars in the second right column). The data support Predictions 2a and 2b and establish Results 2 and 3. Result 2. Global communication reduces effort and increases efficiency when there is no output sharing but effort remains much above optimal levels. Result 3. Global communication tends to increase effort and to reduce efficiency when introduced into the Fixed-Partners or Random-Partners treatments but the effect is small and not statistically significant. Output sharing is somewhat less effective in the presence of Global communication, but continues to achieve near-optimal results. Local Communication and Output Sharing with Linked Fixed and Random Partners colour figure Click here to download line figure Figure 1.docx
doi:10.1007/s10640-017-0124-9 fatcat:cfkjllm5u5et3gqfpu2k3es7te