Common Pool Resource Management and the Effect of Heterogeneous Users: an Experimental Investigation Lance Howe University of Alaska, Anchorage Jim Murphy (future iterations)
Introduction This experiment builds on Walker, Gardner, Ostrom (1990) indirectly by adding heterogeneity (Hackett, Schlager,, Walker (1994)) adding a monitoring and sanctioning institution (Casari and Plott (2003)) adding an endowment switch anticipated and unanticipated Based on 10 controlled experiments (6 at UAA and 4 at UCLA), broad sample of undergraduate students, 80 unique subjects
Introduction Motivation: Decentralization of state forests (local common pool resources) property rights to limited participatory management given to communities India (JFM) and Nepal (Forest User Groups) heterogeneity and the impact of different rules on CPR use and income dist.. Management of CPR s s in rural Alaska: Alaska Native Village Corporation common property ownership arrangements - differing resources, rules and incentives, and users Endowment heterogeneity is an imperfect but useful start to exploring the effect of heterogeneity on CPR management.
Background A foundational CPR experiment (Walker, Gardner, and Ostrom,, 1990) finds in the no-rules environment environment over- investment and near complete rent dissipation the open access outcomes described by G. Hardin In subsequent experiments (Ostrom,, Walker, and Gardner, 1992; Walker, Gardner, Herr, Ostrom,, 2000; Cardenas, Stranland and Willis, 2000) efficiency can be significantly improved through cooperation facilitated mainly by communication and voting Lab experiments support field evidence - small groups of individuals can sustainably manage a common-pool resource with appropriate institutions, and Walker (1994) find that communication is an effective institution even with heterogeneity Hackett, Schlager, and Walker (1994)
Background Existing CPR papers focus on institutions that facilitate coalition building or improved collective action among different users (i.e. common property management of a common pool resource) This experiment explores the effect of monitoring and sanctions alone, with user heterogeneity, in promoting efficient CPR use (key ingredient of CPR management independent of the property right regime)
Background Sanctions in many CPR and Public Goods experiments subjects choose to punish after learning the contributions of other subjects (no monitoring or uncertainty in sanctions) punishment varies & the sanctioning player sets the fine Ostrom,, Walker, Gardner (1992) introduce a sanctioning mechanism to WGO (1990) Fehr and Gachter (2000) punishment is effective in promoting full contributions in public good game. Gurerk, Irlenbusch, Rockenbach (2006) a sanctioning environment is unanimously chosen (with near perfect cooperation) over a sanction free environment in a public goods game no uncertainty in the sanction. Cardenas, Stranlund and Willis (2000) some probability of inspection and a fine concurrent with the over-appropriation (akin to a government managed CPR)
Design This experiment Casari and Plott (2003): a monitoring and sanctioning Institution intended to reflect a local institution used in the Italian Alps for centuries Carte di Regola there exists a pre-announced limit and penalty for over- harvesting uncertain benefits and subjects incur a monitoring cost to inspect other players WGO design adding heterogeneity (Hackett, Schlager,, Walker (1994))
Background Questions/hypotheses explored in current paper 1. Efficiency outcomes are less than predicted efficiency outcomes with heterogeneous endowments 1a. Heterogeneous endowments reduce efficiency outcomes relative to homogeneous endowments 2. The opportunity to monitor and sanction (with weak sanctions) does not improve efficiency outcomes or 2a. A reference point does not change efficiency outcomes. 3. The opportunity to monitor and sanction (with strong sanctions) does not improve efficiency outcomes 4. An endowment switch will not affect efficiency outcomes 5. An unanticipated endowment switch leads to relatively less efficient outcomes compared to anticipated changes. Don t t as yet explore research questions related to monitoring and sanctioning behavior
n x i 1 i x Design In the WGO study, an individuals payoff function is given by: n n 2 [ a( x i ) b ( x i ) ] ( x i ) i 1 i 1 i where x individual investment in CPR i return on outside investment initial endowment
Design 1200 1000 800 600 X 400 200 0 0 10 20 30 40 50 60 70 80 90 100 110 120 130 F( X ) -200 X PO 1 a Nash N a ( ) X ( ) 2 b X open a N 1 b b
Design The payoff function becomes n x i 1 i x i n n 2 [ a( x i) b ( x i ) ] ( k x i) I isi I j i ijr ij i 1 i 1 s H *( x ), for x i i i ri, j ( s j k ), for x j
Design This experiment: Pareto Optimum play 56 total 77 each player Open Access play 112 total low endowment subjects 8 each high endowment subjects 20 each X X PO open 1 a ( ) 2 b a b
Design This experiment: Nash Equilibrium varies by player type low endowment subjects constrained at 8 high endowment subjects x Nash 1 k L i, H ( ) NH 1 b a ( ( b* ) H * p) 16 no sanctions and weak sanctions 14 stronger sanctions (above PO)
Design Sequence of Decision Rounds Heterogeneity Sanctions Exchange Rate Rounds Match 0 No None practice 5 Match 1 Yes None.005 10 Match 2 Yes Weak Sanctions.005 10 Match 3 Yes Stronger Sanctions.005 10 Match 4 Yes (switch endowments) Stronger Sanctions.01 10
Sample Screen-shot stage 1 Design
Sample Screen-shot Stage 2 Design
Treatment # 1: Heterogeneity & no- rules environment environment Predicted Efficiency Outcomes in the No Rules Environment Walker, Gardner, and Ostrom (10 tokens) Walker, Gardner, and Ostrom (25 tokens) Casari and Plott Hackett, Schlager, Walker this Paper (match 1-3) this Paper (match 4) Pareto Optimal Play $ 0.41 $ 0.20 $ 1.22 $ 0.98 $ 0.49 $ 0.98 Nash as % of PO (low end.) 32.7 % 32.7 % 32.7 % Nash as % of PO (high end.) 65.3 % 65.3 % 65.3 % Nash as % of PO (all or avg) 39.5 % 39.5 % 39.5 % 49.0 % 49.0 % 49.0 %
T1: Heterogeneity & no-rules Predicted Efficiency: low Comparing Efficiency Outcomes in the No Rules Environment WGO (10 tokens) WGO (25 tokens) CP HSW 32.7 % Avg. Ind. Efficiency 32.7 % Actual Efficiency: low 35.1 % Predicted Efficiency: high 65.3 % 65.3 % Actual Efficiency: high 84.2 % Predicted Avg. efficiency 39.5 % 39.5 % 39.5 % 49.0 % 49.0 % Actual Avg. efficiency 37.2% -3.16 % 28.4% 45.3%* 59.7 %** *random endowment sessions, no communication ** average individual efficiency over all sessions
T1: Heterogeneity & no-rules Figure 1. Average Investment in the Common Property Resource Across Rounds, Match 1, Low and High Endowment Subjects Mean Public Investment 0 2 4 6 8 10 12 14 16 18 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 Low Endowment High Endowment
T1: Heterogeneity & no-rules Outcome: While low endowment players under-invest a large share of choices are at Nash Public Investment Choices, Match 1 below Nash at Nash above Nash total choices low endowment players 54.8% 45.3% 400 high endowment players 45.8% 7.8% 46.5% 400
T1: Heterogeneity & no-rules Outcome: and a significant percentage of low endowment players play at Nash Public Investment Choices for Low Endowment Players, Match 1 below Nash (8 or more times) at Nash (8 or more times) other strategies subtotal: total players Number of low endowment players 17 13 10 40 Average investment 4.73 7.92 6.76 6.3
T2: Heterogeneity & weak sanctions sanctions Weak sanctions and the effect of a reference point on efficiency outcomes: Match 2 weak sanctions parameter changes don t t affect predicted Nash for low or high endowment players Subjects are provided at least one reference point a a limit of 9 is 9 imposed which players are free to abide by or to ignore Play of low endowment subjects moves closer to predicted Nash
T2: Heterogeneity & weak sanctions sanctions Weak sanctions a reference point?: large number of choices at Nash Distribution of Public Investment Choices, Match 2 low endowment players high endowment players Equal to limit of 9 below Nash (and not = 9) at Nash* above Nash total choices - 29.3% 70.8% - 400 29.5% 32.0% 5.5% 33.0% 400 * Nash equilibrium for low endowment players is equal to 8 and is equal to 16 for high endowment players
T1: Heterogeneity & no-rules Outcome: Information on the limit appears to influence the strategy of low endowment players Public Investment Choices for Low Endowment Players, Match 2 below Nash (8 or more times) at Nash (8 or more times) other strategies subtotal: total players Number of low endowment players 8 24 8 40
T2: Heterogeneity & weak sanctions sanctions Figure 3. Average Investment in the Common Property Resource Across Matches, Low and High Endowment Subjects Mean Public Investment 0 2 4 6 8 10 12 14 16 18 1 2 3 4 1 2 3 4 Low Endowment High Endowment
T2: Heterogeneity & weak sanctions sanctions Efficiency outcomes under weak sanctions: Under weak sanctions, efficiency improves Earnings inequality between subject types declines Actual as a Share of PO (low endowment players (8 tokens)) Actual as a Share of PO (high endowment players (24 tokens)) Actual as a Share of PO (average) Casari Plott Ostrom, Walker Gardner T1 - No Sanctions 35.1 % 84.2 % 59.7 % 28.4%* -3.2% T2 - Weak Sanctions 46.2% 78.5% 62.4% 48.3% 6.5%
T3: Heterogeneity & stronger sanctions sanctions Stronger Sanctions: Under stronger sanctions predicted Nash decreases for high endowment endowment subjects (14) and does not change for low endowment subjects Sanctions are weak relative to Casari and Plott (2003) Low and high endowment players reduce investment in public good
T3: Heterogeneity & stronger sanctions sanctions Figure 3. Average Investment in the Common Property Resource Across Matches, Low and High Endowment Subjects Mean Public Investment 0 2 4 6 8 10 12 14 16 18 1 2 3 4 1 2 3 4 Low Endowment High Endowment
T3: Heterogeneity & stronger sanctions sanctions Share of low endowment players playing at Nash falls to 53 from 71 (still greater than T1) Share of high endowment players at or above Nash nearly constant - 38.5% to 41% Distribution of Public Investment Choices, Match 3 low endowment high endowment at limit of 5 19.3% 26.0% below Nash (and not = 5) 28.0% 32.5% at Nash* 52.8% 2.3% above Nash (and not = 24) 20.0% at 24 19.3% total choices 400 400 average 6.6 12.5 *with stronger sanctions, Nash for low endowment subjects remains at 8 and falls to 14 for high endowment subjects
T3: Heterogeneity & stronger sanctions sanctions Stronger Sanctions: efficiency improvements and earnings inequality declines No Sanctions Weak Sanctions Stronger Sanctions Actual as a Share of PO (low endowment players (8 tokens)) 35.1 % 46.2% 64.2 % Actual as a Share of PO (high endowment players (24 tokens)) 84.2 % 78.5% 88.2% Actual as a Share of PO (average) 59.7 % 62.4% 76.2% Casari Plott 28.4%* 48.3% 76.9% Ostrom, Walker Gardner -3.2% 6.5%
T4: Heterogeneity, Stronger Sanctions and Endowment Switch Endowment Switch: High endowment players increase average investment to levels similar to no- rules environment environment (avg. 12.5 to 15.6) play beyond predicted Nash Low endowment players average investment relatively constant (avg. 6.6 to 6.4) Anticipated (7 sessions) and Unanticipated (3 sessions)
T4: Heterogeneity, Stronger Sanctions and Endowment Switch Figure 3. Average Investment in the Common Property Resource Across Matches, Low and High Endowment Subjects Mean Public Investment 0 2 4 6 8 10 12 14 16 18 1 2 3 4 1 2 3 4 Low Endowment High Endowment
T4: Heterogeneity, Stronger Sanctions and Endowment Switch Endowment Switch: No Sanctions Weak Sanctions Stronger Sanctions Stronger Sanctions Switch Actual as a Share of PO (low endowment players (8 tokens)) 35.1 % 46.2% 64.2 % 47.8% Actual as a Share of PO (high endowment players (24 tokens)) 84.2 % 78.5% 88.2% 71.6% Actual as a Share of PO (average) 59.7 % 62.4% 76.2% 59.7% Casari Plott 28.4%* 48.3% 76.9% Ostrom, Walker Gardner -3.2% 6.5%
T4: Heterogeneity, Stronger Sanctions and Endowment Switch Endowment Switch: With prior information fewer extreme investments (at 24) but about the same number at or above Nash Overall, prior information appears to not to affect efficiency outcomes (or to change efficiency in an unexpected direction) Distribution of Public Investment Choices by Information, High Endowment Players, Match 4 at limit of 5 below Nash (and not = 5) at Nash* 15 to 23 24 total choices No information on endowment switch 11.8% 26.8% 2.5% 28.9% 30.0% 280 Prior information on endowment switch 15.0% 20.0% 4.2% 40.8% 20.0% 120 *with stronger sanctions, Nash for low endowment subjects remains at 8 and falls to 14 for high endowment subjects
T4: Heterogeneity, Stronger Sanctions and Endowment Switch Endowment Switch: Across all treatments, average investment is about equal with and without information for subject initially assigned low endowments Public Investment for players initially assigned low endowments Match 1 Match 2 Match 3 Match 4 Average No Prior Information 6.2 7.3 6.4 15.8 8.9 Prior Information 6.5 7.2 7.0 15.2 9.0
T4: Heterogeneity, Stronger Sanctions and Endowment Switch Endowment Switch: Consequently, efficiency earnings are about the same with greater information for subjects initially assigned low endowments Efficiency outcomes for players initially assigned low endowments Match 1 Match 2 Match 3 Match 4 Average No Prior Information 0.347 0.499 0.649 0.732 0.557 Prior Information 0.360 0.377 0.626 0.680 0.510
Conclusion T1-No Rules Environment: Actual efficiency exceeds predicted Nash efficiency and efficiency outcomes exceed those reported in other studies. Persistent under-investment by low endowment players seems to contribute to this outcome. Investment by low endowment players is bi-modally distributed Consequently, while differences in endowments appear to promote a relatively efficient group outcome overall earnings inequality increases.. Efficiency outcomes are greater than predicted for high endowment subjects and lower than predicted for low endowment subjects in the no no-rules environment.
Conclusion T2 - Weak Sanctions: the opportunity to monitor and sanction improves efficiency,, even when sanctions are designed in such a way as not to affect predicted Nash play. efficiency earnings especially increase for low endowment players and earnings inequality declines. a reference point (i.e. the limit) provided to low endowment subjects may account in part for observed efficiency improvements.
Conclusion T3 - Stronger Sanctions: Similar to results reported in other studies, stronger sanctions (those designed to change predicted Nash equilibrium), significantly improve efficiency. Efficiency earnings improve the most for low endowment subjects and earnings inequality declines further. The improvements we observe are achieved with relatively weak sanctions compared to Casari and Plott.
Conclusion T4 - Endowment Switch: much lower efficiency outcomes with the endowment switch. Even under strong sanctions, a sudden endowment switch results in efficiency outcomes similar to those in the no-rules or open access environment. Investment increases significantly for subjects gaining high endowment positions. additional information on the future endowment switch does not appear improve efficiency outcomes for players initially assigned low endowment positions.
Extensions and Future Research Run additional sessions with information (only 3 information sessions robust??) Directions for additional analysis Analyze monitoring and sanctioning behavior between subject types Explore notions of fairness in sanctioning and investment decisions Future Research Introduce risk to the CPR and or some level of uncertainty to the outside investment option to what extent is the CPR managed to smooth consumption Expand concept of heterogeneity beyond endowments Introduce a choice between types of property regimes