Common Pool Resource with Free Mobility: Experimental Evidence from the US and Mongolia Dolgorsuren Dorj
∗ †
January 28, 2009
Abstract We conducted laboratory experiments in US and Mongolia with the common pool resource setting in which people freely chose between two localities that may differ by governing structures: no regulation or sanctioning mechanism. Governing structures were imposed either exogenously or chosen by majority voting in each locality. We found that (a) under free mobility conditions efficient resource use is attainable if the sanctioning mechanism adjusts to the population level; (b) free riders and cooperators self-select into different regimes; (c) results are consistent across subject pools. Only subtle differences in behavior across countries were observed. JEL classification codes: C7, C70, C91, Q2, R19. Keywords: common pool resource, free mobility, sanctions, experiment ∗
Email:
[email protected] Department of Economics, University of Hawaii at Manoa, 2424 Maile Way, Honolulu, HI 96822. Tel: (808)-948-9093, Fax: (808)-956-4347 † Special thanks go to my advisor, Katerina Sherstyuk for her valuable guidance in this research. I am deeply grateful to Marco Casari, Charles Plott, Lata Gangadharan for helpful comments. Many thanks to Sun-Ki Chai for his effort in establishing the experimental lab at Manoa campus. The experiment was programmed and conducted with the software z-Tree (Fischbacher 2007). This research was funded through the Arts and Science Advisory Council Award at University of Hawaii at Manoa. The research grant from the Department of Economics is greatly appreciated. I would also like to thank conference participants at Western Economic Association International meeting 2008, Economic Science Association North-American meeting 2007.
1
1
Introduction
A number of theoretical and field works explores the problem of resource allocation in common pool resource (CPR hereafter) environments since Hardin (1968). From field studies, Ostrom (1990) identifies successful cases of resource management in the developing world. Casari and Plott (2003) find that villages in the Alps used special monitoring and sanctioning system for managing common resources. Velez, Murthy and Stranlund (2005) in their experiments conducted in the fishing communities of Columbia demonstrate that formal regulation and informal communication may complement each other depending on the local case. Visser (2006) reports increase in cooperation with peer punishment in unequally endowed South African groups. Our work differs from previous research in that it addresses the mobility issue by looking at multiple localities each having CPR problem. Moreover, we compare resource extraction behavior using laboratory experiments across two countries: USA and Mongolia. Our research is motivated by the CPR use in developing countries, such as Mongolia, where the question of overgrazing became most visible in the 1990s. Traditional extensive grazing systems require high mobility and flexibility in the presence of natural disasters and risks. When herders can move freely within a large territory, they are able to manage the risk successfully. The dilemma between risk sharing and the common pool resource problem suggests that with the high-risk weather conditions prevalent in Mongolia, it is advantageous to pool the land into commons and allow herders to move from one common to another. Hence, we test the model based on the notion of community management structure with free mobility. Much of the previous CPR research concerns with the over-exploitation problem within one community. To reflect the high mobility aspect we extend existing research to allow for two communities in which citizens of each area are free to choose the place to appropriate. In each community, there is a land parcel with pastures for grazing activity. We consider two possible community management regimes for each locality: (i) unregulated community with no rules towards the grazing activity; (ii) sanctioning mechanism with mutual monitoring as studied in Casari and Plot (2003) that resembles community property regime. Facing multiple localities with different institutional structures, which community will the herder choose? Particularly, we test the survival of different institutions under the migration pressure condition. Which type of locality will stand the pressure from outsiders when mobility is costless? We address these questions in a multi-community environment CPR framework, where players make decisions regarding the location to graze and the harvesting level. 2
Also, we study whether the endogenous choice of institutions may be supported by the majority voting rule. More specifically, we test whether “voting with the feet” and “voting with the ballot” conditions produce similar results. By “voting by feet” we mean the condition where citizens choose a community with an exogenously given institution while “voting with the ballot” means that citizens vote for the regime by majority rule. The objective of this paper is to study experimentally efficiency of resource use in CPR game with free mobility assuming either no regulation or common property regime within localities. Case studies from Ostrom (1990) suggest that graduated sanctions used against the excessive use of the resource are a helpful tool in CPR environments. However, in a laboratory setting, Walker et. al (1990) report that sanctions do not alter the result of classical Prisoner’s Dilemma game, and the resource is used above the selfish Nash equilibrium because monitoring is a costly activity for the selfish players. In CPR experiments with communication alone, with sanctions alone, or with communication and sanctioning opportunity, Ostrom et. al (1992) find that repeated communication improves average net benefit. They show that with no institutions, resource is destroyed quickly. Gardner et al. (1997) demonstrate that neither entry restrictions nor quota caps increase efficiency above the Nash prediction. Schmitt et al. (2000) examine CPR when only a subset of decision-making group had opportunity of face-to-face communication with each other and find that uncertainty about the outsider’s decision decreases the efficiency of the resource use. Walker et al. (2000) find that proposals and simple majority voting over allocation rules raise efficiency. Experiments by Vyrastekova and van Soest (2003) in contrast to decentralized mechanisms emphasize a centralized enforcement of rules that may be effective only with local community participation by means of majority voting. Casari and Plott (2003) consider a special system “Carte de Regola” that was used to manage the common properties in Alpine villages. They show that revenue-generating sanctioning structure among villagers produces high level of cooperation. A decentralized sanctioning system of “Carte de Regola” has an advantage over other sanctioning systems because it reduces the deadweight loss to the society. Fines imposed on violators are not a pure cost to the system; it is a revenue (transfer) to monitors. In addition, sanctions are imposed on the violator only once, so there are no multiple fines on the same action. The only loss in a system of mutual monitoring is the inspection cost. Also, the punishment level is not subjective; it depends on the deviation of players from the allowed level (threshold) and is proportional to the amount of excessive use. In this experimental study we focus on the classical CPR model and consider the sanctioning mechanism of Casari and Plott (2003). Our contribution is to add 3
free mobility to the model. In this paper our aim is to (i) study whether under free mobility, individual’s behavior changes in response to changes in the resource governing structure; (ii) consider effects of exogenous vs. endogenous institutions on resource use under free mobility conditions and (iii) test whether the sanctioning system may improve the welfare across two localities. Our experiments show that: (a) under free mobility conditions, efficient resource use is still attainable in communities which adopt sanctions; (b) subjects “vote with the feet” for sanctioning system over unregulated regime so that more people locate in the regulated (sanctioning) locality; (c) partial monitoring is sufficient to induce higher level of efficiency in sanctioning regime. These results are explained by the majority of subjects who are rationally responded to institutional incentives and cooperated in a regulated locality, but overused the resource in an unregulated community. We call such people “conditional cooperators”. At the same time we find that there are significant minorities who are less sensitive to institutions. We call “free riders” those subjects who tend to over-harvest irrespectively of institution, and “unconditional cooperators” those who restrain harvesting at or below socially optimal level regardless of institution. Regarding these behavioral types, we find that (d) there is a sorting of population: free riders cluster in the unregulated locality and overuse the resource; in contrast, cooperators gather in the regulated locality keeping efficiency at high level; (e) subjects “vote with the ballot” for and against sanctioning system depending on their behavioral types; cooperators vote for sanctions most of the time whereas free riders vote for no regulation at the beginning, but eventually through competition learn that sanctions are a better governing structure. (f) With respect to subject pool, the results are consistent across US and Mongolia subjects. However, the voting decisions by behavioral types have subtle differences across subject pools. The next section provides predictions of CPR free mobility model. Section 3 explains experiment design. Section 4 reports the results of the experiment with and without free mobility. In section 5 we discuss and conclude.
2
Theoretical Predictions
We use a standard CPR model (e.g. Falk, Fehr and Fischbacher 2002) and the sanctioning mechanism of Casari and Plott (2003) outlined in section 2.2. We assume that all subjects are self-regarding, identical and that the social benefit of adopting sanctions outweighs the social cost. Therefore, adopting the sanctioning mechanism improves welfare. We assume that a community with sanctions adjusts 4
threshold to the changed population level in the case of migration entry or exit. Further, if institutions are endogenous, we assume the majority voting rule determines the institutions. We use terms “harvesting”, “grazing” and “appropriation” interchangeably throughout the paper.
2.1
One locality, No Sanctions
The baseline common-pool resource game assumes no regulation. A finite number of herders n, each with endowment e, simultaneously decide on the amount of appropriation from the common pool denoted as xi . Here i denotes the index for herder. P Let X = ni=1 xi be the sum of all herder’s appropriation and f (X) be total revenue of all herders, where f (X) = aX − bX 2 is a concave function with parameters a > 0 and b > 0. The cost of maintaining one animal denoted by c is common and independent of every other member’s decision while the revenue for each member will depend on grazing choice of all members. The socially optimal appropriation level is given by X opt = (a − c)/2b and it is independent of the number of people in the locality. Hence, in a symmetric socially optimum outcome, each herder harvests xopt = X opt /n. The share of individual i in total appropriation is xi /X. Then each herder i0 s profit is given by πi = e − cxi + [xi /X]f (X). The symmetric Nash equilibrium appropriation by each herder (e.g. Falk, Fehr and Fischbacher 2002) is characterized as follows: Proposition 1 In a symmetric Nash equilibrium with homogeneous herders the apa−c . The total appropriation in Nash propriation by each herder is given by: x∗i = b(n+1) n a−c ∗ equilibrium is given by X = n+1 · b , which is higher than the social optimum, Xsopt = (a − c)/2b, if n > 1.
2.2
One locality, Sanctions
The model with sanctions (Casari and Plott 2003) assumes that users of the resource set up rules among themselves in such a way that, the Nash equilibrium total appropriation in the locality is at the socially optimal level. In particular, the community agrees and restricts appropriation to certain threshold amount per herder, λ. All herders are free to monitor each other. Harvesting beyond this threshold costs individual a fine payment if any other member of community discovers his/her violation. For each excess harvesting unit a violator pays unitary fine, h. The fine the violator pays is a transfer to the inspector who discovers the violation. Monitoring is a costly decision for anyone who decides to inspect any other member. 5
By paying inspection fee, k, the inspector may obtain exact information about the harvesting decision of one other member. The payoff for each herder is given by P πi = xXi f (X) + e − cxi − Ii mi + j6=i Iij rij , where the total fine paid by violator, mi = h(xi − λ) with xi > λ, generates revenue, rij = mj − k, to herder i who monitors herder j. Iij = 1 if herder i inspects herder j, Iij = 0 otherwise. Note that inspection will take place if herder i violates the rules and some other herder j inspects i. If the herder i has been monitored by more than one inspector, j 6= i, a person randomly selected out from requested herders receives the direct transfer P from monitoring. Thus, Ii = 1 if j6=i Iij ≥ 1, and Ii = 0 otherwise. Note that there are no multiple fine payments for violators. If more than one inspection is requested for a given person, one of herders will be randomly selected as an inspector. The following is established in Casari and Plott (2003): Proposition 2 Suppose a locality has a sanctioning mechanism, where the threshold and punishment levels are set as λ = xopt − k/h − ε, where ε > 0 is small enough and h = a − c − xopt (n + 1) · b, respectively. Then this sanctioning mechanism supports the socially optimal level of harvesting at the Nash equilibrium, X ∗ = X opt . Also in this equilibrium everyone inspects each other, with inspection probability being equal to one, p∗ = 1. Efficiency level in community is equal to 100 percent if the monitoring cost equals zero 1 . However, it is less than 100 percent if we account for the monitoring cost.
2.3
Sanctions in one locality and No sanctions in the other locality
In order to obtain predictions with free mobility across two localities we consider a Tiebout-type economy (1956), where each locality faces a CPR problem rather than a local public good problem. We use a Nash type free mobility equilibrium concept, a partition of population that is stable against any unilateral deviations to existing jurisdictions 2 . Hence, with homogeneous herders in equilibrium, all Q∗ Qopt Q Qopt Q∗ Efficiency is defined as: E = / ·100 ( −n · e)/ ·100. Here is the sum of Q= opt indicates maximum social surplus. One can actual profits minus sum of endowment while also say sanctions restore efficiency, yet not without a loss, a monitoring cost is a waste of the resources. We characterize an outcome in terms of efficiency as an optimality of appropriation level and overall efficiency, which includes a monitoring cost. Efficiency and an overall efficiency coincide when a monitoring cost is equal to zero. 2 Two equilibrium concepts, Nash equilibrium and the core of coalition structure, are used in local public good models (see Greenberg and Weber 1986). The Nash equilibrium is defined as the 1
6
localities are inhabited; no citizen wants to move to any other jurisdiction, and profits are equalized across localities. Suppose there are two communities of herding families located in neighboring areas denoted by Si ∈ {1, 2}. We allow free mobility across the herding communities, which means that migration is a costless decision. All herders are perfectly mobile between the two localities, but can live in only one locality at a time. From the herder’s point of view, both communities are identical except for their local regulatory mechanism, which all herders take as given. Consider two neighboring communities where one locality has sanctions and the other locality has no regulation at all. First, herders choose a locality. Once the number of herders in each locality is known, the locality with sanctions announces the threshold. Second, each herder decides on a grazing level. Third, in the locality with sanctions, herders may monitor each other after observing the total appropriation. All discovered violations become a public information. Recall that in free mobility equilibrium all localities are inhabited and no citizen wants to move to any other locality. Hence, the equilibrium condition results in equal individual profits across localities. Proposition 3 Assume the regulated locality can optimally adjust its harvesting threshold to the population level. Then (1) The total appropriation level in the community with sanctions, XR , will remain at the socially optimal level and equal to XR = X opt . Efficiency of the sanctioning system is at maximum. (2) Efficiency of the unregulated pasture improves as compared to the case where there is no sanctioning system neighboring the unregulated pasture. Here, migration to the regulated locality reduces the population in the unregulated locality, and efficiency increases; yet appropriation is above socially optimal level, XU > X opt . (3) The appropriapartition of agents into jurisdictions, where no single agent wants to move from the current position to join other existing jurisdiction. That is, the concept of Nash equilibrium, often referred to as a free mobility equilibrium or Tiebout equilibrium, does not allow any unilateral deviation by agents (Westhoff 1977, Greenberg 1983, Epple et al. 1984, 1993, Dunz 1989, Konishi 1996, Nechyba 1997). The core of coalition structure (Strong Tiebout Nash equilibrium, coalition-proof Nash equilibria) is an alternative interpretation of the Tiebout’s equilibrium, where no group of agents wants to form a new jurisdiction or join existing one to make members better off (see McGuire 1974, Wooders 1978, Berglas and Pines 1981). We use the first concept, Nash equilibrium, in our two-community CPR model, for the following reasons. First, theoretically the core may not exist (Wooders 1980). Second, in terms of applications, we are interested in extensive animal production, which is an individual family based activity that does not require group formation yielding economies of scale. Third, the distance between families and lack of a comprehensive communication tool make the group formation unavailable. Fourth, seasonal migration from one place to other prevalent in extensive production stands as an obstacle to any stable coalition structure.
7
tion levels across localities are as follows, X opt = XR < XU . The locality with the sanctioning system has a higher number of individuals than the unregulated locality, NR > N U 3 . Note that sanctions in one locality create a positive externality for the unregulated neighboring locality, and efficiency increases in both localities as compared to the no regulation case.
2.4
Voting Equilibrium in localities with free mobility
In this part we relax the assumption of exogenous institutions. We allow each community to choose their own governing institution using majority voting (Fiorina and Plott 1978). We show that the regulatory regime can be sustained in both localities in a subgame perfect Nash equilibrium under majority voting. Again, each individual chooses first a locality. Next, each member of the community votes either for sanctions or no sanctions in their locality, and the outcome is determined by majority voting. Third, each herder decides on an individual harvesting level. In the communities with sanctioning regime, monitoring decisions follow. Proposition 4 In the voting equilibrium, agents vote for sanctions in both localities. The resource is used at the efficient level. The appropriation levels across localities are the same, XR1 = XR2 = X opt and the population sizes are identical, NR1 = NR2 . This result is obtained using the median voter theorem (Duncan 1948, Downs 1957). Since in our model all agents are identical and share homogeneous preferences, by the median voter theorem the outcome of the majority voting in each locality is the median voter’s preferred institution, that is, sanctions.
3
Experimental Procedures and Design
Computerized experiments of the CPR game with free mobility involved 40 undergraduates and graduate students from the University of Hawaii at Manoa (USA) and 80 students from the Academy of Management (Mongolia) in May-June 2007. Ten participants were invited to each session. Subjects were seated separately from each other at laboratory computer terminals. The instructions were provided in verbal and written format. See the instructions in the Appendix of paper. No communication was allowed. The CPR problem was described as an abstract decision-making 3
See Dorj (2007) for detailed proofs.
8
situation, where there exists an opportunity to make a money by “investing” in market(s). Each of the ten subjects chose investment level without knowledge of the others’ investments. The investment level was expressed in “tokens” and payoffs were in terms of “francs” 4 . Each subject was paid in cash at the end of the session. Subjects received 5 dollars for participation and earned on average additional 17.5 dollars in the USA (US hereafter) site, while the average payoff in Mongolia (MG hereafter) site was 12.5 dollars. Sessions lasted less than two hours. We used the following parameters for the production function: a = 14.5, b = 1/30, c = 2.5, N = 10 in total, n = 5 per locality (market). In the locality with sanctions, the monitoring cost was set at k = 7 and the threshold λ and fines, h, varied with the population level 5 . Each subject had 10 tokens per period as endowment. The range of tokens that the subjects were allowed to order was [1, 500]. Each token that the subject ordered cost him c = 2.5 experimental francs. Gross group return and return on tokens invested were presented in Table 1 of instructions (part 1, N). Also each subject was provided with a more detailed table, where total group investment increased by single unit increments. We conducted four treatments in two designs (See Tables 1 and 2). In each session, several treatments were implemented sequentially. Design 1 included three treatments: no mobility and no sanctions (N -treatment ), followed by no mobility and sanctions (S-treatment), followed by free mobility and asymmetric institutions (N S-treatment), where one market had no regulation and other market had sanctions. The first two treatments investigated the effect of the institution (sanctions) on the efficiency of resource use. The last treatment (N S-treatment) tested the effect of free mobility on the institution (sanctions) performance. Note that the institutions here are exogenous, and subjects in the NS-treatment“vote with the feet”, choosing their preferred locality in a Tiebout sense. The second design had four treatments: no mobility and no sanctions (N), followed by no mobility and sanctions (S), followed by free mobility and asymmetric institutions (NS), followed by free mobility and voting (V). In addition to the three treatments described above, in V-treatment subjects voted for an institution and made investment decision after 4
Laboratory artificial currency that was converted into domestic currency, either US dollars or Mongolian tugrugs, at the end of each session. 5 See Table 2 in Part 3 of Instructions. From Proposition 2 a threshold (λ) was adjusted to the population size as follows: no threshold if n=1 ; λ=87 if n=2; λ=58 if n=3; 43 if n=4; 34 if n=5; 28 if n=6; 24 if n=7; 21 if n=8; 18 if n=9; 16 if n=10. From Proposition 2 a fine (h) varied with the population size in the following way h=0 if n=1 ; h=3 if n=2; 4 if n=3; 4.5 if n=4; 4.8 if n=5; 5 if n=6; 5.14 if n=7; 5.25 if n=8; 5.33 if n=9; 5.4 if n=10. See Casari (2005) on fine-to-fee ratio.
9
choosing the market. There were two markets labeled “A” and “B”, and subjects initially chose one of the two markets. Having observed the number of participants in his/her market, each subject voted for an institution. After the voting procedure in each locality, the chosen institution was announced. The voting treatment captured the differences between “voting with the feet” and “voting with the ballot” conditions. Within-subjects design was chosen to let the subjects experience both N and S conditions before the two-market free mobility NS-condition. The sequence of treatments replicated historical realities, where tragedy of commons have been experienced first (N-treatment), and the regulatory mechanism often followed as a solution to the free rider problem (S-treatment). Also, we pursued a structure that allowed subjects to start with simple instructions before going on to more complex environments. We used “locality”, “community”, and “market” interchangeably, since in the experiment the locality for grazing activity was presented to subjects as a market for investment. A subject’s decision consisted of the following actions: (i) choice of a market in the free mobility treatments only; (ii) voting decision in the voting treatment only; (iii) investment decision; (iv) inspection decision in markets with sanctions. In our experiments we employ within-subject design in order to see how individual behavior changed with institutions. Instructions were read aloud to everyone at the beginning of each treatment. After the instructions were read, a quiz was given before each treatment to ensure that they understood the instructions clearly. Subjects had two practice periods to become familiar with the software. At the end of the session a questionnaire was administrated which provided feedback on the strategy the subjects followed. Each period, the computer screen displayed a history table of the previous plays within subject’s own group only (for one market treatments) or within both markets (for free mobility treatments). We now describe specific procedures for each treatment.
No sanctions in one market (N -treatment) In N -treatment, ten subjects were randomly placed in two groups of five subjects each at the beginning of the session and were told they will remain in the same group throughout the duration of the treatment. The partner-matching reflects long term interaction among herding families in one community. In each period, each participant had an opportunity to invest between 1 to 500 tokens in their market. Subjects were informed about the patterns of return in the market which depended on the total investment. There were eight rounds of play with feedback on 10
total investment, return per token in the market, and own payoff after each round.
Sanctions in one market (S-treatment) The same group of subjects as in N-treatment then experienced the sanctions treatment, where the threshold was announced. After an investment decisions were made, each participant could inspect others whom they suspected of having violated the common threshold level. Monitoring per person required 7 francs, while each token discovered above the threshold transferred 4.8 francs from violator to the inspector. If many people asked to monitor a particular person, one of them was randomly chosen as an inspector; all others were treated if they did not monitor and did not incur the cost of seven francs. After each round, participants were informed about the total investment and return per token in the market, their payoff from investment, monitoring revenue if they inspected others, and fine paid if they violated and were inspected by others. Again, a group of five made up one market for the eight periods.
Free mobility and Asymmetric institution (N S-treatment) The free mobility treatment was modeled by giving each of the ten subjects a choice to invest either into unregulated market A or regulated market B. At the beginning of each round, subjects first chose the market in which to invest. After the number of people in each market was known, the threshold λ was announced in the regulated market B, and the remainder of the round proceeded as in N-treatment in the unregulated market A and as in S-treatment in the regulated market B. The threshold λ in the B-market was adjusted given the number of subjects in the particular market. See Table 2 in the NS-instructions (part 3). Fines also varied with population size in the market with sanctions. After each round, participants received feedback on the total investment and return per token in each market, their payoff, and inspection results in the regulated market. Extra tokens placed by a violator who was monitored appeared next the violator’s ID and were publicly observable to all ten subjects. We did not change the labeling of the two markets across rounds and treatments in order to help coordination; market A stayed unregulated, and market B stayed regulated for the duration of the NS-treatment.
11
Free Mobility and Voting by Ballot (V-treatment) In the voting treatment, subjects were informed that regulation in each market will be determined by the majority voting rule among the subjects in the particular market. In the case of a tie, the regime was chosen randomly by the computer. Subjects first decided which market to select from two markets labeled “A” and “B”. Second, after having observed the number of people in the market, they voted for sanctions or no sanctions in their chosen market. Once the regime was decided by voting, either sanctions or no sanctions treatments were implemented. After each period, subjects in both markets received feedback on the number of participants in each market, voting outcome in each market, total investment in each market, return per token in each market, their payoff and violation level in each market if sanctions were the outcome. The theoretical predictions for each treatment, given the parameter values, are displayed in Table 1.
4
Results
There are in total 12 sessions, four in the USA and eight in Mongolia. The summary of the experimental sessions is provided in Table 2. We analyze the data from each site, US and Mongolia, separately and pooled together. Overall results in both sites are very similar except for the voting stage behavior. Hence, below we present results for the pooled data. Occasionally, we compare both sites data when it is necessary to distinguish the differences that have arisen. Table 3 provides summary statistics for the pooled data and each separate site. First, we analyze the data in terms of average behavior and compare them with predicted values for each treatment. We evaluate performance of institutions by group use, individual use, efficiency, population split with free mobility, equilibrium profits in two localities, monitoring under sanctions, and voting outcome. Finally, we move the analysis to the individual level and see how individual behavior varies with institutions.
4.1
No mobility and No Sanctions
We use the classical Nash model in one locality as a benchmark and compare it with the experimental results where no sanctions were introduced. We employ Wilcoxon signed ranks test for matched-pairs to compare actual values with predicted values within each treatment. Our within-subject design allows to use the same test to
12
compare results across treatments. Therefore, we report p-values for the Wilcoxon signed ranks test, unless indicated otherwise. Result 1 With no sanctions, the resource extraction is above the socially optimal level. Efficiency of the resource use is low and no different from the Nash equilibrium prediction. The mean group use and the mean individual use were slightly below the predicted Nash level. Support (Table 3, Row N ). Group Use: The overall average group appropriation for the twelve experiments was 280.7, as compared to social optimum of 180, however below the Nash level of 300 at a 5 percent significance level (p=0.0186). 88 percent of data falls into 25 percent bandwidth around the predictions. Individual use: Out of 640 individual decisions, 13.8 percent was exactly at Nash equilibrium (60 tokens). Overall, 57 percent of the data is within the 25 percent bandwidth (i.e. in the interval [45, 75]) around the predicted value. 14.1 percent of individual actions was above 75 tokens, and 28.9 percent of actions was below 45 tokens. Individual use was slightly below the Nash predicted value (p=0.0186). Efficiency: Groups were able to absorb only 61.5 percent of the surplus available in the market, which was not significantly different from the Nash equilibrium prediction (p=0.0597).
4.2
No mobility and Sanctions
With sanctions the Nash model predicts no overgrazing. We use the terms “monitoring”, “regulation” and “sanctions” interchangeably. Result 2 With sanctions total appropriation drops dramatically and group efficiency improves substantially. Average of the group resource use and average of the individual use were slightly above the Nash equilibrium predictions. Support (Table 3, Row S.) Group Use: Average group use dropped from 280.7 in the unregulated locality to 193.3 with sanctions. It was slightly above the social optimum of 180 in the data pooled across subject pools (p=0.0121), whereas for the separate subjects pool data the group use was no different from the socially optimal level at 5 percent significance level (p-values are 0.0687 and 0.0679 respectively for MG and US sites). 13
Individual use: The mean individual use dropped from 56.1 tokens without sanctions to 38.7 tokens with sanctions and statistically was no different from the Nash predictions of 36 for both sites ( p-values are 0.0678 and 0.0679 respectively). In total 65.5 percent of all actions were at or below the Nash equilibrium prediction. Efficiency: In terms of efficiency, overall surplus from resource use reached on average 94 percent and it is below the predicted value, 96.8, at one percent significance level (p= 0.0037). However, there is a huge difference between the surplus absorbed in the no sanctions and the sanctions treatments (61.5 vs. 94 percent, differ at any significance level). Examination of two institutions in one locality produced contrasting outcomes, where with no regulation the resource is overused while sanctions prevents overextraction. Therefore, we confirm that, as in Casari and Plott (2003), sanctioning system prevents the overgrazing problem of CPR.
4.3
Free Mobility and Asymmetric institution (NS-treatment)
Now we analyze the data in the two-locality design with free mobility, where one locality (labeled A) was unregulated and the other locality (labeled B) had sanctions. Result 3 Under free mobility condition, the community with sanctions sustains a much higher resource use efficiency than the unregulated community. Group use in the unregulated community is higher than in the regulated community, and no different than the Nash equilibrium prediction. As predicted, more people locate in the community with sanctions. Support (Table 3, Row NS) Group Use: The locality with sanctions experienced an overall low average group use at 191.8. In contrast, the average group use in the unregulated locality across twelve experiments was 280.2. Group use in the unregulated community was statistically no different from the predicted Nash value at 5 percent level (p=0.3465). Results from the regulated community were slightly above the social optimum (p=0.0186, two-tailed). As predicted in terms of direction, the appropriation in the unregulated locality was significantly higher than the resource use in the regulated locality at 1 percent level (p=0.0022). Efficiency: As predicted, efficiency in the locality with regulation was higher compared to the unregulated regime. In the locality with sanctions, the mean efficiency across twelve sessions was 93.1, which is slightly below the socially optimal 14
level. Average efficiency in the unregulated locality ranged from minimum of −220 to maximum of 100 6 . The mean was 47.9 for the pooled data and below the prediction. We will discuss later how observed self-selection and competition among free riders in the unregulated locality drove down the efficiency. Population split: The predicted split of population, 4 : 6, across the localities is recorded in 23.9 percent of data. Overall 62.5 percent of the data were within the 25 percent bandwidth (i.e. in the interval [3, 5] and [5, 8]) around the predicted value. The locality with sanctions had larger population than the unregulated locality (p=0.0025). If we look at the consistency of participation in each markets, in each session there were at least six to ten subjects who chose the same market 90 percent of the time. The results indicate that in the “voting with the feet” condition, institutions matter in the sense that a much higher level efficiency ia obtained with sanctions than without sanctions. Equilibrium forces put the localities in a balance such that the majority of people chose the regulated market “B”, and fewer people invested in the unregulated market “A”. However, unlike the theoretical prediction (Proposition 3) we did not observe a positive externality from the regulated locality on the unregulated locality. Efficiency of the unregulated locality in asymmetric institution NS-treatment was in fact lower than in the locality with NN-treatment (47.9 vs. 61.5 percent). As we will show below in section 4.7, this was due to clustering of free riders in the unregulated locality. This clustering of free riders in the unregulated market “A” and their predatory behavior washes out the benefit of neighboring with regulated community so that efficiency in the unregulated locality is no higher than in the one-locality no mobility case.
4.4
Free Mobility and Voting by Ballot
Recall that in this two-locality design, institutions are formed endogenously, e.g. each period subjects choose the locality for the appropriation, then subjects choose the regime by majority vote within the chosen locality. When the regime is announced, subjects decide on the individual appropriation level. If sanctions is the voting outcome, subjects may monitor others after observing the total appropriation within the group. 6
Efficiency is negative if the return per token in this locality is negative.
15
Result 4 In the majority of cases, voting resulted in the sanctioning regime. When sanctions were chosen the resource use was at the efficient level as predicted; when no sanctions were chosen, the resource was over-used. Support (Table 3, Row VV, Figure 1a-1f.) Voting Outcome: Recall that according to the prediction, subjects suppose to vote for sanctions in 100 percent of the cases. The data shows that in 64 percent of the time, sanctions were implemented as a result of voting across sessions. Two types of voting dynamics are emerged. In 3 out of 6 sessions (session 7, 8, and 10) there was a convergence to sanctions in both localities such that in market “B” sanctions was the only outcome in all periods, and market “A” gradually arrived to the regulated regime with the initial play of no sanctions. However, in two other sessions (6 and 9) we observed a divergence of regimes: in market “A” no regulation dominated and in market “B” sanctions prevailed. Session 5 in MG site exhibited unsettled patterns preventing any type of classification. See Figures 1a-1f. The data shows that the mean percent of the population who voted for the S-regime was significantly lower in market “A” than in market “B” in both sites (p=0.0069 in MG and p=0.0166 in US). This again suggests sorting of types into distinct markets. In terms of dynamics across periods, the average percent of people who voted for sanctions in market “A” was far less than in market “B” in the first period (36 vs. 74). However, gradually, the percent of people voting for sanctions rose from the fourth period, and by the seventh period, the average number of votes against regulation had dropped and votes for sanctions were about the same in both localities. Use level: In market “B” the Nash prediction of 180 is confirmed (p=0.1159) while average group appropriation in market “A” was statistically different from the Nash equilibrium level (p=0.0277). However, calculations based solely upon the regime type resulted in average group uses that were no different from the Nash equilibrium levels. With sanctioning system within the locality, individual use was much smaller at 36.8 tokens than with no sanctions at 63.9 tokens (Wilcoxon ranksum test p-value=0.0039). Efficiency in the no sanctions regime was different from the sanctions regime at 1 percent significance level (p=0.0039). Population split: The observed population split closely reflected the predicted split of 5:5 across localities. Overall 43 percent of the data are within the 20 percent bandwidth (i.e. in the interval [4, 6]) around the predicted value.
16
With voting, communities chose sanctions as a governing regime in a majority of the cases . In line with the predictions, the sanctioning mechanism, if chosen, sustains efficient resource use. As predicted, about half of the subjects appropriated in one locality and another half appropriated in the other locality. At the same time, voting on the appropriate regime might not immediately bring efficiency. Most of the time market “A” implemented no regulation while market “B” predominantly chose sanctions regime 7 . The efficiency in the locality labeled “A” was lower than that in locality “B”(p=0.0464). Mean efficiency in market “B” was no different from predicted efficiency in the S-treatment (p=0.9165). This suggests that in voting condition there was sorting of population into distinct markets. As it will later be discussed in section 4.7, subjects used market labeling as a coordinating device such that followers of sanctions clustered in market B and frequently voted for sanctions, quickly obtaining high efficiency. In contrast, free riders clustered in the market “A” and implemented mostly the unregulated regime.
4.5
Monitoring
We now look at whether, under the sanctioning regime, monitoring decisions (inspections) were in line with prediction. Result 5 Theory predicts one hundred percent of inspections such that entire population will be inspected. In contrast, partial monitoring was sufficient to induce higher level of efficiency. Support (Table 3) Monitoring: In the S-treatment 74.4 percent of actions were inspected. In the NStreatment with asymmetric localities and free mobility 81 percent of all actions were inspected in the locality with sanctions, and inspection rates were no different from S-treatment rates (p=0.4802). In the voting treatments with free mobility on average 75.3 percent of all actions were inspected in localities with sanctioning regime. Therefore, approximately three quarter of the population were inspected in the markets with sanctions. Monitoring was less than the predicted 100 percent. On average 11.6 percent of maximum surplus was transferred as fines from violators to 7
Overall in 53 percent of cases one locality had no regulation and the other market had sanctions where 43 percent of time locality labeled “A” had no regulation whereas locality “B” had no regulation in only ten percent of time. Out of seventy cases 26 (37 percent) were situation where both localities had sanctions, and in ten percent of cases both localities ended up with no regulation.
17
inspectors. Average monitoring cost amounting to 2.6 percent of maximum surplus was smaller than the predicted value of 3.2 percent. In terms of cost administering the institution, experiments demonstrate that subjects do not monitor when people do not violate. With sanctioning regime, partial monitoring was sufficient to increase the efficiency of the resource use.
4.6
Tiebout equilibrium hypothesis
Now we test Tiebout equilibrium hypothesis in the two-localities case. Recall that the free mobility equilibrium is characterized by partition of the population where no citizen wants to move and where profits are equalized in both localities. Result 6 In the free mobility treatment with asymmetric institutions (NS-treatment), the average earnings in the unregulated locality were lower than in the regulated locality. Similarly, in the voting V-treatment the locality with sanctions had higher average per capita earnings than the locality without sanctions. Support Individual profits: The average per capita profits in the two localities with exogenous institutions (NS-treatment) were statistically different from each other at 1 percent level (p=0.0076). The average per capita profit in the regulated locality, 179, was no different from predicted level of 173 (p=0.1823), while the average profit in the unregulated locality, 133, was below the Nash equilibrium level at 1 percent significance level (p=0.0096). Wilcoxon matched-pair testing shows that in two the localities with endogenous institutions (V-treatment) the average profits were different at 5 percent significance level (p-value=0.0464). Again, across the twelve sessions in locality “B”, where sanctions were the dominant regime, average per person profit, 197, was no different from equilibrium level (p=0.1712), while the average profit in locality “A”, 163, was below Nash predicted level of 173 at 5 percent significance level (p=0.0215). Period by period data on difference in per capita profits also does not show any convergence to zero toward the end of sessions. Average period profits were higher in the regulated market sixty percent of all periods. We believe that these results are due to self-selection resulting free riders to cluster in market “A” and cooperators in market “B”.
18
4.7
Individual Behavior
Here we analyze the data on the individual level. We define “cooperative appropriation” as a harvesting action which is at or below the socially optimal level. On the other hand, “free riding” will be referred to an action which is above the social optimum, xopt . Result 7 Individual actions changed with institutions: the cooperative actions were prevalent under the sanctioning regime while free-riding actions were prevalent under no regulation. Support (Table 4, Figure 2) Institution effect: As predicted, individual behavior changed consistently with treatments. This confirms that institutions critically affect and shape individual behavior. Percent of cooperative actions was 19.3 in (N), 20.1 in (NS) and 19 in (V) treatments with no regulation, as compared to 65.5, 73.5, and 70.7 percent respectively in (S), (NS), and (V) treatments with sanctions. Results in both sites, MG and US, reveal the same pattern in individual behavior. We further classify behavioral types based on actions taken by participants in the no mobility N-treatment and S-treatment. We define a “cooperator” as a person who complies with the socially optimal investment level more than fifty percent of the time in both the N-treatment and S treatments. “Free riders” exceed harvesting levels in both N and S treatments in more than half of their real choices. “Conditional cooperators” free ride in N-treatment and comply with the rules in S-treatment in more than half of the cases. Result 8 Heterogeneity among subjects is documented. The majority of subjects were conditional cooperators; however, the proportions of unconditional cooperators and free riders were also non-negligible. Support (Table 5). Heterogeneity: Overall, the data shows that 12.5 percent of the population were unconditional cooperators, 33.3 percent were free riders. The majority of the subjects (51.7 percent) were conditional cooperators; their behavior dependent upon the institution. From a total of 120 subjects, only three had behavior inconsistent
19
with either of the 3 categories 8 . See Table 5. Composition of types stays the same with the classification of behavioral types based on the single period action 9 . Result 9 Labeling of the markets such as “A” and “B” served as a coordinating device for the market choice decision. In a majority of the cases cooperators and conditional cooperators clustered in market “B” with sanctions, and free riders went to market “A” that had no regulation. Support (Figure 3). Sorting: Let’s look again at the treatments where citizens are free to choose localities. Note that we use here terms “market” and “locality” interchangeably. In the NStreatment, unconditional cooperators and conditional cooperators on average went to the regulated market (S) in majority of time while free riders mostly joined the unregulated locality (N). From Figure 3 we can see that free riders on average chose regulation 43 percent of the time while conditional cooperators in both MG and US sites went to the regulated market 67 and 54 percent of the time respectively. The NS-treatments in MG site show that free riders went to the regulated market significantly less often than conditional cooperators (Wilcoxon-Mann-Whitney ranks sum test p=0.0004). However, in the US site there was no statistical difference between percent of time conditional cooperators joined the regulated market and percent of time free riders spent in the regulated market (Wilcoxon-Mann-Whitney ranks sum test p=0.619). In both sites unconditional cooperators and conditional cooperators spent the same amount of time in the regulated B market (WilcoxonMann-Whitney ranks sum test p=0.4744 and p=0.0619 respectively). Subjects in the US site used market labeling from the NS-treatment as a focal point for their decision to join one of the two markets in the consecutive voting treatments. Recall that we used the same labeling (“A” and “B”) in V-treatment as in NS-treatment. This labeling helped unconditional and conditional cooperators to join market “B”, expecting sanctions (S), and free riders to cluster in market “A”, expecting it to have no sanctions (N). In the voting treatments, conditional cooperators in the US site went to the B market more often than free riders (Wilcoxon8
See also Houser and Kurzban (2005), Gunnthorsdottir et al. (2007), Isaak and Walker (1988), where the cut-points were respectively 50, 30 or 33 percent of endowment in the public good context or classifications in Casari and Plott (2003) for spiteful, altruistic and selfish motives. 9 In terms of evolution of types, one third of the population in the NS-treatment and half of the population in the V-treatment changed their behavior from their types defined based on (N and S)-treatments classification. We believe this is explained by the complexity of two-locality conditions in both exogenous and endogenous institutions setting.
20
Mann-Whitney test p=0.0226). In contrast, in MG site the free riders spent the majority of their time in the B market as conditional cooperators (Wilcoxon-MannWhitney test p=0.1733). In the US site 67 percent of the time no regulation was the outcome in market “A”, whereas 94 percent of the time sanctions were implemented in market “B”. This was due to the sorting of free riders in market “A” and cooperators in market “B”. Hence, in the US site, sanctions were implemented less often in market “A”, whereas in MG site, sanctions were implemented in both markets (52 percent in “A” and 75 percent in “B”). Therefore the sorting behavior was easily observable in MG site with exogenous institutions and in US site with endogenous institutions. In the MG site with voting, there was no clear cut between clustering behaviors among free riders and conditional cooperators. Both types in MG site spent most of the time under the S-regime (sanctions) in the V-treatments. However, in the US site with voting, there was a definite sorting of conditional cooperators into the market with sanctions and free riders into no regulation market. It appears that in both the exogenous (NS-treatment) and the endogenous (Vtreatment) settings, availability of two simultaneous markets leads to population sorting into distinct communities. In market B, cooperators clustered and efficiency was high. Next we will see how in market A competition among free riders reduced efficiency of the resource use. Result 10 In the locality where free riders clustered, efficiency was low until the final rounds of play. The free rider’s “push out others” strategy was not successful and vanished toward the end of the sessions. Free riders eventually learned that the “push out others” strategy is wasteful. Support (Table 4, Figure 1c-1f ). Competition among free riders: Besides the sorting of the population, we also observed an interesting phenomena called “push out others”. The more opportunistic subjects in the unregulated market were harvesting huge amounts and suffered current losses in an attempt to earn more surplus the next period by persuading concurrent competitors leave the market. There were five cases where one or two subjects extracted excessive amounts, resulting in market losses for everyone (e.g. unregulated locality in NS-treatment in the sessions 3, 4, 8, 12; V-treatment of the session 9). These losses caused the majority to leave the market. This behavior is similar to the predatory pricing behavior common in competitive industries (Tirole 21
1988). However, with some experience, free-riders realized the presence of more than one subject demanding monopolist position. Hence, “push out others” strategy was wasteful, leading to reductions in surplus. Toward the end of the session, in the market where free riders cluster (sessions 7, 8, 10), there was a convergence to sanctions regime in both markets (see Figure 1c, 1d, 1f ). Non-compliant subjects in the unregulated locality learned that with sanctions they could earn more profit than without sanctions. Institutions taught them how to behave (Li and Plott 2007; Brown, Kamp and Plott 2007). Now we look at differences across types in monitoring and voting behavior. Result 11 Overall, the conditional cooperators monitored more than other types. At the same time, if sanctioning was the voting outcome in the locality labeled “A”, then free riders were more engaged in monitoring activity. Support (Table 3, Figure 5). Monitoring by behavioral types: The percent of inspections done by conditional cooperators was higher than the percent of monitoring by unconditional cooperators at 5 percent significance level (p=0.0206). At the same time, in V-treatment, if sanctioning was the voting outcome in locality labeled “A”, then percent of monitoring in market “A” was significantly higher compared to the monitoring in the “B” market that had sanctions (90.5 vs. 68.8 , p=0.0350)). Recall that market “A” was mostly occupied by free riders. Hence, in this market, free riders highly monitored others if sanctions regime was the voting outcome. This result is similar to Casari and Plott (2003) finding that free riders like to monitor others, and the presence of free riders helps turn spiteful motives into socially desired outcomes. In our free mobility experiments, due to self-selection, free riders had less of a chance to have sanctions and monitoring outcome because their modal choice of institution was often no regulation. Majority of the inspections was done by conditional cooperators, about 1/5 of inspections was implemented by free riders, and the least monitoring was done by unconditional cooperators. Result 12 Free riders tend to vote for no regulation, while cooperators and conditional cooperators vote for sanctions. Support (Figure 6). Voting: In the voting treatment both localities could choose sanctions, however this did not happen from the beginning. There was a sorting of population according 22
to their types which resulted in the implementation of sanctions 80 percent of the time in market “B”, and no sanctions 53 percent of the time in market “A”. In US site free riders on average voted for regulation 35 percent of the time, while unconditional and conditional cooperators voted for sanctions 66 and 70 percent of the time respectively. In US site, conditional cooperators voted for S-regime more often than free riders (Wilcoxon-Mann-Whitney test p-value=0.0351). However, in MG site, free riders and conditional cooperators voted for monitoring 64 and 69 percent of the time respectively. Wilcoxon-Mann-Whitney test shows no difference between the voting behaviors of conditional cooperators and of free riders (p-value=0.5399) in MG site. See Figure 6 for difference in the free riders behaviors in the two sites. To distinguish between the effect of the institutions and the player’s inherent nature on subject’s behavior, we ran a regression where the dependent variable was the harvesting choice and the independent variables included the realized regime and the subject’s choice of regime. The implemented regime and voting choice both had significant effect on the harvesting level (p=0.000, p=0.002); with the change from no sanctions to sanctions regime, the harvesting level decreased by 22 tokens. The realized regime (standardized coefficient on the OLS regression, beta=0.49) by majority votes had more relative strength than the subjects’ voting choices itself (beta=0.12). In the voting treatments, the divergent behavior among behavioral types was most evident in US site, meaning that the percent of time conditional cooperators voted for the S-regime was higher than the percent of time free riders voted for the S-regime. Also, percent of time spent in the resulting regulated market was higher for conditional cooperators than for free riders in US site. In comparison, in MG site there was no difference in the voting behavior among types; both conditional cooperators and free riders voted mostly for the sanctions and ended up in the regulated market. Therefore, behavioral type sorting was more observable in endogenous setting in US site, while in MG site sorting of types into different regimes was more evident in the exogenous institutions setting. This suggests that subjects in MG site had less coordination in the V-treatment and didn’t use the labeling from the NS-treatment for their decision to join markets.
5
Discussion
This paper reports the results of an experiment that studied CPR problem with multiple localities. In the one-locality and two-locality designs, behavior is compared to point predictions for the symmetric Nash equilibrium and the socially optimal 23
outcome. We constructed four treatments to test the predictions of the free mobility model. In particular, with no regulation, the community overexploited it’s resources beyond the socially optimal level and efficiency was significantly lower than efficiency in the locality with sanctions. Our one-locality treatment results are in line with the Casari and Plott (2003) finding that sanctioning leads to improved efficiency of CPR use. Interesting results are derived from the treatment where one locality adopts sanctions and the other does not have regulation. As predicted, the sanctioning system keeps the resource use at an efficient level. Unregulated locality has dropped efficiency to the level below the predicted value. Overall, the lower mean efficiency in the unregulated locality was mostly due to the regime. In the voting treatment, subjects in the locality labeled “B” voted for sanctions most of the time and efficiency was high, no different from predicted level; however subjects in the locality labeled “A” appropriated above the socially optimal level. Because sanctioning was implemented less often in the “A” market than in “B”, the use level was higher in “A” than in “B”. However, results based upon the realized regime were in line with the predictions. Our analysis of individual level data reveals heterogeneity among population. In the two-locality treatments, the clustering of non-cooperative types into the unregulated market and their attempt to establish monopolist position drove down the efficiency. Thus, the sorting of population into free riders in the market “A” and both conditional and unconditional cooperators in the market “B” brought patterns that were not predicted by the free mobility equilibrium. Interestingly, predatory behavior and rivalry among free riders lowered the efficiency in the unregulated market. Furthermore, we find that behavioral types differed in their voting and monitoring decisions in the following way: free riders voted for no sanctions most of the time while cooperators preferred sanctions. Overall conditional cooperators monitored more than the other behavioral types. However, if the sanctions were implemented in market “A” then percent of monitoring by free riders was highest. Interestingly, in the voting treatments, subjects used market labeling as a coordinating device. With endogenous institutions, again the clustering of free riders in the “A” market and their votes for no sanctions lowered efficiency below the predictions. However, with multiple experience free riders realized eventually that voting for sanctions is a better way to improve their payoffs, and after several periods or at least by the end of session, equilibrium converged to the sanctions regime in half of the sessions. “Voting with the feet” and “voting with the ballot” conditions produced similar results by sorting out types into two distinct communities. However, comparison across two conditions reveals a short-term advantage of the former construct in 24
terms of policy design. In the sanctions condition where the institution was given in advance, the welfare increased from the very beginning as compared to no sanctions. In the “voting with the ballot” condition subjects ended up with the social dilemma unresolved, or learned after a few periods that sanctioning is the better way to sustain the resource. For policy makers, this suggests that the local participatory decision may take time to establish the appropriate institution. The results obtained above are robust across the subject pools except subtle differences in behavior across types. In V-treatments, conditional cooperators in US site went to the regulated market more often than free riders and voted for sanctions more often than free riders; while in MG site both types voted for sanctions, resulting in S-regime in a majority of the cases. Therefore, less coordination and less sorting was evident in MG site than in US site during the V-treatment with endogenous setting. However, sorting was more evident in MG site than in US site in the NStreatment with exogenous institutions. Also subjects in the US site demonstrated stable voting pattern throughout all sessions whereas subjects in one of the session at the MG site showed unsettled pattern in voting behavior. All of the above differences may reflect first, complexity of voting treatment; second, a lack of experience in the voting and local participation by the subjects in MG site. Our results draw several conclusions. Most importantly, we find that the sanctioning institution under the free mobility condition may survive if the harvesting adjusts to the migration process. In our experiments with exogenous institutions, welfare increased from the beginning, while with endogenous institutions, it took time to establish the appropriate regime. Partial monitoring was sufficient to obtain high efficiency with sanctions. The presence of multiple-localities may bring about the sorting of subjects according their behavioral types such that cooperative types cluster in the community with sanctions and non-cooperative types in unregulated locality. This might be one reason why some regions are slow in success while others quickly progress.
References [1] Berglas, E., Pines, D., 1981. Clubs, local public goods and transportation models: a synthesis. Journal of Public Economics 15, 141-162. [2] Brown, A. L., Kamp, H., Plott, C. R., 2007. The Nature of collusion facilitating and collusion breaking power of simultaneous ascending price and simultaneous descending price auctions. Economic Inquiry, forthcoming. 25
[3] Casari, M., Plott, C. R., 2003. Decentralized management of common property resources: experiments with a centuries-old institution. Journal of Economic Behavior and Organization 51, 217-247. [4] Casari, M., 2005. On the design of peer punishment experiments. Experimental Economics 8, 107-115. [5] Dorj, D., 2007. Common pool resource with free mobility: voting with their herds. Mimeo [6] Downs, A., 1957. An Economic Theory of Democracy. Harper Collins. [7] Duncan, B., 1948. On the rationale of group decision-making. Journal of Political Economy 56, 23-34. [8] Dunz, K., 1989. Some comments on majority rule equilibrium in local public goods economies. Journal of Economic Theory 47, 228-234. [9] Epple, D., Filimon, R., Romer, T., 1984. Equilibrium among local jurisdictions: toward an integrated treatment of voting and residential choice. Journal of Public Economics 24, 281-308. [10] Epple, D., Filimon, R., Romer, T., 1993. Existence of voting and housing equilibrium in a system of communities with property taxes. Regional Science and Urban Economics 23, 585-610. [11] Falk A., Fehr, E., Fischbacher, U., 2002. Appropriating the commons: a theoretical explanation. In: Ostrom, E., Dietz, T., Nives Dolsak, Stern, P. C., Stonich, S., Weber E. U. (eds.) Drama of the Commons. Committee on the Human Dimensions of Global Change, National Research Council, National Academy Press, 157-192. [12] Fiorina, M. P., Plott, C. R., 1978. Committe decisions under majority rule: an experimental study. The American Political Science Review 72(2), 575-598. [13] Gardner, R., Moore, M., Walker, J., 1997. Governing a groundwater commons: a strategic and laboratory analysis of western water law. Economics Inquiry 35(2), 218-34. [14] Greenberg, J., 1983. Local public goods with mobility: existence and optimality of a general equilibrium. Journal of Economic Theory 30, 17-33. 26
[15] Greenberg, J., Weber, S., 1986. Strong Tiebout equilibrium under restricted preference domain. Journal of Economic Theory 38, 101-11. [16] Gurerk, O., Irlenbusch, B., Rockenbach, B., 2006. The competitive advantage of sanctioning instiutions. Science 312, 108-110. [17] Gunnthorsdottir, A., Houser, D., McCabe, K., Ameden, H., 2007. Disposition, history, and contributions in public goods. Journal of Economic Behavior and Organization 62, 304-315. [18] Hardin, G., 1968. The Tragedy of the Commons. Science 162, 1243-48. [19] Isaac, M., Walker, J., 1988. Group size eEffects in VCM provision: the voluntary contributions mechanism. The Quarterly Journal of Economics 103, 179-199. [20] Konishi, H., 1996. Voting with ballots and feet: existence of equilibrium in a local public good economy. Journal of Economic Theory 68, 480-509. [21] Kurzban, R., Houser, D., 2005. Experiments investigating cooperative types in humans: a complement to evolutionary theory and simulations. PNAS 102(5), 1803-1807. [22] McGuire, M., 1974. Group segragation and optimal jurisdictions. Journal of Political Economy 82, 112-132. [23] Nechyba, T., 1997. Existence of equilibrium and stratification in local and hierarchical tiebout economies with property taxes and voting. Economic Theory 10, 277-304. [24] Ostrom E., 1990. Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press, New York. [25] Ostrom, E., Walker, J., Gadner, R., 1992. Covenants with and without a sword: self-governance is possible. American Political Science Review 86, 404-417. [26] Plott, C. R., Li, J., 2005. Tacit collusion in auctions and condition for its facilitation and prevention: equilibrium selection in laboratory experimental markets. California Institute of Technology, Social Science Working Paper, 1223. [27] Schmitt, P., Kurtis J. S., Walker, J. M., 2000. Collective action with incomplete commitment: experimental evidence. Southern Economic Journal 66 (4), 829854. 27
[28] Tiebout, C., 1956. A pure theory of local expenditure. Journal of Political Economy 64, 416-424. [29] Tirole, J., 1988. The Theory of Industrial Organization, MIT Press. [30] Visser, M., 2006. Welfare implications of peer punishment in unequal societies. Working Papers in Economics 218, Goteborg University, Department of Economics. [31] Velez, M. A., Stranlund, J. K., Murthy, J. J., 2005. What motivates common pool resource users? Experimental evidence from the field. Journal of Economic Behavior and Organization, forthcoming [32] Vyrastekova, J., Soest, D., 2003. Centralized common-pool management and local community participation. Land Economics 79 (4), 500-514. [33] Walker, J., Gadner, R., Ostrom, E., 1990. Rent dissipation in a limited-access common-pool resource: experimental evidence. Journal of Environmental Economics and Management 19, 203-211. [34] Walker, J., Gadner, R., 1992. Probabilistic destruction of common-pool resources: experimental evidence. The Economic Journal 102 (September), 11491161. [35] Walker, J.M., Gardner, R., Herr, A., Ostrom, E., 2000. Collective choice in the commons: experimental results on proposed allocation rules and votes. The Economic Journal 110, 212-234. [36] Westhoff, F., 1977. Existence of Equilibria in Economies with a Local Public Good. Journal of Economic Theory 14, 84-112. [37] Wooders, M. H., 1978. Equilibria, the core, and jurisdiction structures in economies with a local public good. Journal of Economic Theory 18, 328-348. [38] Wooders, M. H., 1980. The Tiebout Hypothesis: Near optimality in local public good economies. Econometrica 48 (6), 1467-1486.
28
Table 1: Equilibrium Predictions Social N ash M ax Eqlm Ef f − cy Optimum Eqlm Surplus Surplus percent 5 X opt = 180 X = 300 Π = 1080 Π = 600 55.6 opt x = 36 x = 60 π = 120 (S) 5 X opt = 180 X = 180 Π = 1080 Π = 1045 96.8 P P opt x = 36 x = 36 = π− k π = 209 p(monitor) = 1 opt (N,S) 10 Xi = 180 X(N )A = 288 Π = 1080 Π(N )A = 691.2 64.0 (loc.A) i = A, B X(S)B = 180 per loc. Π(S)B = 1038 96.1 (loc.B) P P nA = 4 = π− k nB = 6 x(N )A = 72 x(S)B = 30 π(N )A = 172.8 π(S)B = 173 p(SB , monitor) = 1 (V) 10 Xiopt = 180 X(A) = 180 Π = 1080 ΠA = 1045 96.8 i = A, B X(B) = 180 per loc. ΠB = 1045 nA = 5 per loc. nB = 5 xA = 36 xB = 36 πA = 209 πB = 209 regimeA − sanctions regimeB − sanctions p(SA , monitor) = 1 p(SB , monitor) = 1 N- number of subjects; Eqlm-equilibrium; Eff-cy -efficiency (N)-No sanctions; (S)Sanctions; (NS)-No sanctions in one and Sanctions in other market; (V)-Voting; (A)-labeling for market with no sanctions in (NS)-treatment; (B)-label for the market with sanctions in (NS)-treatment; (A)-label for market 1 in (V)-treatment; (B)-label for market 2 in (V)-treatment; π-per capita profit; Π-surplus excluding endowments; X-total group appropriation, x-individual appropriation; k-per capita monitoring cost; p(monitor)-probability of monitoring. T reat ment (N)
N
29
Table 2: Summary of Experimental Sessions Number of Date Session Code Site periods Design1 : N, S, N S (N ), (S), (N, S) 8, 8, 16 May 21, 2007 11 US 8, 8, 16 May 24, 2007 12 US 8, 8, 16 June 5, 2007 1 MG 8, 8, 16 June 7, 2007 2 MG 8, 8, 16 June 8, 2007 3 MG 8, 8, 16 June 9, 2007 4 MG Design2 : N, S, N S, V (N ), (S), (N, S), (V ) 8, 8, 6, 8 May 22, 2007 9 US 8, 8, 6, 10 May 25, 2007 10 US 8, 8, 6, 14 June 6, 2007 5 MG 8, 8, 6, 10 June 7, 2007 6 MG 8, 8, 6, 14 June 8, 2007 7 MG 8, 8, 6, 14 June 9, 2007 8 MG US-University of Hawaii at Manoa, Honolulu, USA ; MG-Academy of Management, Ulaanbaatar, Mongolia .
30
Table 3. Mean efficiency, group use, individual use, population, and inspection.
Mean Average efficiency, %
(stdev)
MG
US
Pre di cted
All
MG
US
61.5
61.3
61.8
300
280.7*
282.1
277.9
(13.7)
(17.7)
(4.6)
(20.5)
(25.2)
(13.7)
94*
93.6*
94.8
180
193.3*
191.5
196.8
(3.2)
(3.7)
(2.3)
(15.5)
(15.5)
(11.5)
47. 9*
47
49.7
280.2
276.8
286.9
(21.1)
(25.5)
(16.2)
(23.1)
(27.1)
(18.2)
96.1
93.1*
93.3
92.9
180
191.8*
189.7
196.1
(3.1)
(3.5)
(3.2)
(12.7)
(13.7)
(13.2)
96.8
72.8*
78.8
69.5
180
246.7*
238.9
249.9
(15.6)
(4.3)
(18.4)
(17)
(11.7)
(17.5)
90.9
85.1
97.5
(9.7)
(12.9)
(2)
Treatment
Predi cted
(N)
55.6
(S)
96.8
(NS) N
64
S (VV) VA
VB
96.8
Regime N
55.6
(VV) frequency,%
Actual
All
59.1
67.3
42.6
(25.5)
(5.3)
(2.6)
0
36
36
36
96.8
94.2*
93.3
96.2
(41.5)
(3.7)
(1.7)
64
64
64
p-value:** N=S
0.0022
0.0117
p-value:** N=S in NS
0.0022
p-value:** VA=VB
Regime S frequency,%
100
N
S
p-value:** V =V
Average Group appropriation (tokens)
288
180 288
180
Actual
199.3
205
180.2
(30.8)
(40.8)
(10.8)
276.5
270.1
289.3
(11.9)
(5.9)
(11)
Average individual use (tokens) Predi cted 60
Actual
All 56.1*
US 55.6
~
~
~
~
~
~
~
~
~
~
100
74.4*
4
4.3*
4.3
4.36
~
~
(0.4)
(0.4)
(0.2)
5.7*
5.7
5.6
100
80.9*
(0.4)
(0.4)
(0.2)
5
5.1
4.4
(0.4)
(0.4)
(0.7)
(5)
(2.7)
38.2
39.4
(2.8)
(3.1)
(2.3)
72
65.4*
65.2
65.8
(8.84)
(13.6)
(3.4)
33.9*
33.4*
34.8
(2.59)
(2.1)
(2)
49.5*
47.3
56.8
(4.24)
(4.2)
(4)
60
36
Act Ual All
56.4
38.7*
36
Predi cted
MG
(4.29)
36
Actual
Percent of actions inspected, %
Predi cted
36
30
Average # of people per locality
40.5
42.1
32.2
(9.57)
(5.7)
(7.1)
63.9
64.7
62.2
(23.8)
(9.7)
(12.1)
All
MG
US
(16.2)
6 5
5
5
4.9
5.6
(0.4)
(0.4)
(0.7)
4.6
4.5
5.0
(0.7)
(0.8)
(0.6)
5
5.5*
5.7*
5.2
(0.4)
(0.3)
(0.3)
5
(14.3) 100
90.5* (7.5)
100
68.8* (18.8)
~
100
~
191.9
193.9
188.2
36.8
36.4
37.7
(15.4)
(21.4)
(3.7)
(12.8)
(1.9)
(0.4)
75.3*
0.0679
0.0022
0.0117
0.0679
0.0022
0.0117
0.0679
~
~
~
~
0.0117
0.0679
0.0022
0.0117
0.0679
0.0022
0.0117
0.0679
0.0025
0.0140
0.0679
~
0.0464
0.1441
0.1797
0.0464
0.1441
0.1797
0.1730
0.4652
0.1797
0.9156
0.5807
0.6547
0.0350
0.0277
0.0679
0.1797
0.0277
0.0679
0.1797
0.0277
0.0679
0.1797
0.0747
0.0679
0.6547
~
(13.6)
*-different from predicted value at 5 percent level in Wilcoxon signed ranks test statistics that compare actual results with predicted value; standard deviations are in parentheses; **Wilcoxon matched-pairs signed ranks test, two sided; N- no sanctions; S-sanctions; NS-no sanctions in one and sanctions in other market; VV-voting in two markets, VA-voting in A market, VB-voting in B market, VN- N-regime in voting, VS- S-regime in voting.
1
Regime
Action
Table 4. Percent of cooperative actions by institution (N vs. S) MG site US site Treatment N S All N N 19.1 ~ 19.1 19.6 S ~ 65.8 65.8 ~ NS 23.6 76.1 53.8 13.4 By regime, V All 18.9 69.8 53.3 19.1 loc. A 21.4 55 40.1 19 loc. B 14 81 66.1 20 Total 6.2 47.1 53.3 6.7 % of time each regime was chosen in V-treatment All 36 64 100 36 loc. A 48 52 100 67 loc. B 25 75 100 6
S ~ 65 68.3
All 19.6 65 44.1
All pooled data N S 19.3 ~ ~ 65.5 20.1 73.5
All 19.3 65.5 50.6
73.5 62.1 77.3 47.8
54.4 33.3 74.2 54.4
19 20.6 14.5 6.3
70.7 56.1 79.9 47.3
53.6 38.9 68.3 53.6
64 33 94
100 100 100
36 53 20
64 47 80
100 100 100
Cooperative action refers to harvesting at or below social optimal level; loc.-locality, N-no sanctions, S-sanctions, NS-sanctions in one and no sanctions in other locality, V-voting treatment
2
Table 5. Composition of individuals by behavioral types, % Treatment (N), (S)
Behavioral Types
All, %
# subjects
Unconditional cooperators Free riders Conditional cooperators Irrational
12.5 33.3 51.7 2.5
15 40 62 3 120
MG site All, % subject, # 13 34 51 3
10 27 41 2 80
1
2
3
4
1 3 6 0
2 0 8 0
2 3 4 1
3 3 3 1
Session 5 6 2 6 2 0
0 2 8 0
7
8
0 6 4 0
0 4 6 0
US site All, % subject, # 13 33 53 3
5 13 21 1 40
9
10
11
12
4 2 4 0
1 3 6 0
0 4 6 0
0 4 5 1
Cooperators: cooperate >50% in both (N) & (S) Free riders: free ride >50% in both (N) & (S) Rational: free ride >50% in (N) & cooperate >50% in (S) Irrational: cooperate >50% in (N) & free ride >50% in (S)
3
Figure 1a. Voting outcom e: regim e 1-sanctions, regim e 2-no santions, (MG site) session 5.
Figure 1b. Voting outcom e: regim e 1-sanctions, regim e 2-no sanctions, (MG site) session 6. 3
2
A market
1
B market
Regime
Regime
3
2
A market
1
B market
0
0 1
2
3
4
5
6
7
8
9 10 11 12 13 14
1
2
3
4
5
Period
8
9
10
Figure 1d. Voting outcom e: regim e 1-sanctions, regim e 2-no sanctions, (MG site) session 8. 3
2
A market
1
B market
Regime
3 Regime
7
Period
Figure 1c. Voting outcom e: regim e 1-sanctions, regim e 2-no sanctions, (MG site) session 7.
2
A market
1
B market
0
0 1
2
3
4
5
6
7
8
1
9 10 11 12 13 14
2
3
4
5
6
7
8
9 10 11 12 13 14
Period
Period
Figure 1e. Voting outcom e: regim e 1-sanctions, regim e 2-no sanctions, (US site) session 9.
Figure 1f. Voting outcom e: regim e 1-sanctions, regim e 2-no sanctions, (US site) session 10.
3
3
2
A market
1
B market
Regime
Regime
6
2
A market
1
B market
0
0 1
2
3
4
5
Period
6
7
8
1
2
3
4
5
6
Period
7
8
9
10
Figure 2. Percent of cooperative harvesting actions by treatment 100 Percent
80
76
66 65
70 74
68
60 40 20
MG site 24
19 20
13
19 19
0 N
S
N in NS S in NS N in V S in V Treatment Cooperative action refers the harvesting at or below social optimal use, results are robust with the threshold as a cut-point
US site
Figure 3. Mean percent of times each behavioral type went to regulated market 90
100 Percent
80 60
67 52
43
54
65 43
40
69
68
68 38
32
20 0 MG-S in NS
US- S in NS
MG-S in V
Treatment Cooperator
Conditional cooperator
Free rider
US-S in V
Figure 4. Inspections by behavioral types and treatment 100
Percent
80 60 40 20
24
32 20
26
29
22
13
31
26
28 16
7
0 MG-NS
US-NS
MG-V
US-V
Treatment Cooperator
Conditional cooperator
Free rider
Percent
Figure 5. Mean percent of times each behavioral types vote for S-regime in V-treatment
100 80 60 40 20 0
69
66
54
70 35
14 MG
US Treatment
Cooperator
Conditional cooperator
Free rider
INSTRUCTIONS (N)
instructions for no sanctions condition in a single market
This is an experiment in decision-making. The instructions are simple and if you follow the instructions carefully and make good decisions you may earn a considerable amount of money. Your earnings and $5 show up fee will be paid in CASH in private at the end of the experiment. Conversion rate is $______ per 1 experimental franc. You are NOT allowed to communicate with any other participant. From this point onwards, you will be referred to by your participant number. Your ID number is at the left top corner of the screen. You will make money by INVESTING tokens in a market that will give you a cash return for your tokens (See example #1 below). The experiment in which you are participating is comprised of several parts, each having a sequence of periods. In each period you will be asked to make an investment in a MARKET. PART 1: INVESTING In this part of the experiment, you will be in a market four other participants. What happens in your market has no effect on the people in the other market, and visa versa. You will not be told which people are in your market, but you will be in the market with the same four other people during Part 1 of the experiment. In each period you can invest a number of tokens between 0 and 500 in the market. All other participants can also invest up to 500 tokens and so the total group investment is at most 2500 (= 500 times 5 people). For every token you invest, you will be charged 2.5 francs, and you will be collecting your return in experimental francs from the market. The return from the market depends on the number of tokens you invest as well as the amount all others in the group invest. The total group investment determines the gross group return (See attached table A) and you will receive a fraction of it according to your personal investment level. Your earning depends on total group investment, gross group return, number of tokens you invest, cost, and endowment in a following way
Your Earnings=
Gross Group Re turn x Your Investment - Cost + Endowment Total Investment Your Share of Gross Group Return
The example below explains the computation in detail. You will make your decision before knowing other people’s investment decisions in that period. You are not to reveal your investment decision to anyone. Other participants’ decisions are private to you; however you will be informed about the total group investment and your profit at the end of each period. EXAMPLE #1 Suppose you invest 20 tokens in the market and the rest of participants invest in total 200 tokens. The cost of your tokens is 50 francs: Cost of your Investment= 2.5 francs per token x 20 tokens = 50 francs In this example, the total group investment is 220 tokens (20 tokens you invest plus 200 tokens by others). The corresponding gross group return is 1577 francs, as shown in the table A attached. The first column of the table lists the total group investment and the second column gives you the corresponding gross group return. The last column indicates return per token invested and obtained by dividing gross group return (in 2nd column) to the total group investment (in 1st column).
1
You will receive a share of the gross group return. For your share of gross, you can multiply your personal investment level by the “Return per token invested” column of the table A, that is 20 *7.17 =143.4. Your net return is 93.4 francs (your share of gross, 143.4, minus the cost of the tokens, 50). Finally, your earning for each period is the sum of your net return and endowment that will be given to you in advance. The period endowment is a constant amount and does not depend on the investment decisions. To find your period earning add values in columns (8) and (9) of the following table. To sum up, in this example your period earnings in francs are given by (assume you are ID #5): ID #
Number of token invested (your investme -nt)
Total Group Invest ment
Gross Group Return
Return per token
(table A)
(table A)
(1)
(2)
(3)
(4)
5
20
220
1577
Your Share of Gross
(5)=(4)/(3)
(6)=(2)x(5)
7.17
143.4
Cost
Net return
(7)=2.5 x (2)
(8)=(6)-(7)
-50
93.4
Endo wme nt
Period Earning
(9)
(10)=(8)+(9)
+10
103.4
The gross group return is graphed below.
Gross Group Return (francs)
Gross Group Return on Market 1600 1450 1300 1150 1000 850 700 550 400 250 100 -50 -200 0 -350
48
96
14
4
2 19
24
0
28
8
33
6
4 38
43
2
48
0
52
8
6 57
62
4
67
2
72
0
8 76
81
6
86
4
91
2
0 96
Total Group Invetsment (tokens)
Notice that the gross group return on the market can be negative if the total group investment is sufficiently large. For instance, if each person invests 110 tokens, the total group investment is 550 tokens and the gross group return is –200 francs. When considering the cost of the tokens, each person has to pay 315 francs. In each period you may record your investment and earnings information in the record sheets provided. This investment decision will continue for a number of periods. Your earnings for each period will be added to find your cumulative earnings for this part of the experiment. If you have any questions concerning the instructions feel free to raise your hand and an instructor will assist you. Please, go through the review question in the next page and fill in the blank lines with the values you think are correct.
2
PART 1_N
PRACTICE ID ______
REVIEW Consider the following investment decisions: ID# 1 2 3 4 5 Total group investment tokens 40 40 40 60 60 240 Suppose you are person #1. To compute your net return on the market, take your investment of 40 tokens and multiply it by 6.5 (Return per token invested, second column of the table A) = 260 francs is your share of gross. Your net return is 160 francs (your share of gross = 260 minus the cost of tokens = 100). Now, you go on and fill in the blank lines for ID#2 and ID#5. ID #
Number of token invested
(1)
(2)
1 2 5
40 40 60
Total Group Invest ment
Gross Group Return
Return per token
(table A)
(table A)
(3)
(4)
240
1560
(5)=(4)/(3)
6.5
Your Share of Gross (6)=(2)x(5)
260
Cost
Net return
(7)=2.5 x (2)
(8)=(6)-(7)
-100 -
160
Endo wme nt
Period Earning
(9)
(10)=(8)+(9)
+10 +10 +10
170
Please, raise your hand if you have any questions and an instructor will assist you.
At the beginning we will run a practice-period experiment to get familiar with the rules. It will NOT count towards your earnings.
ARE THERE ANY QUESTIONS?
3
Table A. GROSS GROUP RETURN Total group investment (1)
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400 410 420 430 440
Gross Group Return (2)
0 142 277 405 527 642 750 852 947 1035 1117 1192 1260 1322 1377 1425 1467 1502 1530 1552 1567 1575 1577 1572 1560 1542 1517 1485 1447 1402 1350 1292 1227 1155 1077 992 900 802 697 585 467 342 210 72 -50
Return per token invested (3)=(2)/( 1)
0.00 14.17 13.83 13.50 13.17 12.83 12.50 12.17 11.83 11.50 11.17 10.83 10.50 10.17 9.83 9.50 9.17 8.83 8.50 8.17 7.83 7.50 7.17 6.83 6.50 6.17 5.83 5.50 5.17 4.83 4.50 4.17 3.83 3.50 3.17 2.83 2.50 2.17 1.83 1.50 1.17 0.83 0.50 0.17 -0.11
Total group investment (1)
450 460 470 480 490 500 510 520 530 540 550 560 570 580 590 600 610 620 630 640 650 660 670 680 690 700 710 720 730 740 750 760 770 780 790 800 810 820 830 840 850 860 870 880 890
Gross Group Return (2)
-116 -152 -173 -185 -192 -195 -197 -198 -199 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200
Return per token invested (3)=(2)/( 1)
-0.26 -0.33 -0.37 -0.39 -0.39 -0.39 -0.39 -0.38 -0.38 -0.37 -0.36 -0.36 -0.35 -0.34 -0.34 -0.33 -0.33 -0.32 -0.32 -0.31 -0.31 -0.30 -0.30 -0.29 -0.29 -0.29 -0.28 -0.28 -0.27 -0.27 -0.27 -0.26 -0.26 -0.26 -0.25 -0.25 -0.25 -0.24 -0.24 -0.24 -0.24 -0.23 -0.23 -0.23 -0.22
Total group investment (1)
900 910 920 930 940 950 960 970 980 990 1000
Gross Group Return (2)
-200 -200 -200 -200 -200 -200 -200 -200 -200 -200 -200
Return per token invested (3)=(2)/( 1)
-0.22 -0.22 -0.22 -0.22 -0.21 -0.21 -0.21 -0.21 -0.20 -0.20 -0.20
1. Find out the actual group investment 2. Look at the average return of each token 3. Multiply the average return by the number of tokens you have invested
4
INSTRUCTIONS (S) instructions for the sanctions condition in one market PART 2: In this part of the experiment you will make your money by INVESTING them in the market that will give you cash return for your tokens as in a previous part of the experiment. Also you may earn money by MONITORING other people’s decisions and eventually getting some revenues from the inspections. Each period consists of two stages: Investment and Monitoring. The investment procedure will be exactly the same as in the PART 1 of the experiment. After the investment stage you will have chance to earn additional revenue from monitoring others decisions. MONITORING After the total group investment is revealed, you will have the chance to impose a payment to the people that invested more than 34 tokens in the market. Notice that if the total group investment is more than 170 tokens (that is 34 tokens times 5 people), at least one person invested more then 34 tokens. You don’t know the individual investments of the other people, but you can ask to uncover them by paying 7 francs for every person you ask to inspect (inspection fee). If the person inspected invested more than 34 tokens, she pays 4.8 franc for every extra token. You get this money (inspection revenue) and everybody will know the investment level of the person inspected. You make the monitoring decision when you know the total group investment, but before knowing other people’s monitoring decisions. An identification number will be assigned to every person to maintain anonymity and it must be considered strictly confidential. EXAMPLE #2 Suppose the total group investment is 280 tokens and your investment is 40 tokens (you are ID #1). Before the monitoring you know that 110 extra tokens were invested (=280 -170) but you don’t know who invested them. Well, you did part of the job with 6 tokens (=40 – 34), but there are other four people around that invested 104 extra tokens. Suppose you ask to inspect person #2 and she has invested 64 tokens. You pay an inspection fee of 7 francs and get an inspection revenue of 144 francs (= (64 – 34) * 4.8) where 4.8 is the fine per each extra token invested. After your inspection, everybody will know that pers#2 has invested 64 tokens, but your identity will not be revealed. In sum, besides the period earnings from investment already explained above, your earnings will be affected by your and other people’s monitoring decisions as follows: Monito EndowPeriod Cost of Monito Fine if Your Total ID Number Inspecte ring ment Earning tokens -ring Share Group # of token d cost revenu of Invest invested e Gross ment (1)
(2)
(3)
1 2
40 64
280
(4)
(5)=2.5 x (2)
206.8 330.8
-100 -160
(6)=[(2)34]x4.8
(7)
(8)
+144 ~
~ -144
-7 ~
(9)
+10 +10
(10)=(4)(5)+(6)-(7)(8)+(9)
253.8 36.8
If you (ID#1) get inspected, then you will pay for every token above 34 and this fine would be (4034) x 4.8=28.8. If two or more people ask to inspect the same person, only one inspection will be executed. A person will be randomly selected by computer and she will pay the inspection fee and get the eventual inspection revenue. The other inspectors will be treated as if they did not ask to inspect that person. ARE THERE ANY QUESTIONS?
5
INSTRUCTIONS (N-S) PART 3:
instructions for the two-asymmetric market free mobility condition
In this part of experiment you will make your money by INVESTING them in the market that will give you cash return for your tokens. There are two markets: MARKET A and MARKET B. You can only invest into one market in any period. If you are investing in MARKET B you may earn money by both investing and monitoring other people’s decisions. MARKET A has investment and no monitoring. In each period you will be asked to make either an investment in MARKET A or an investment and a monitoring in MARKET B decision. There are NINE other participants in this experiment who will be asked to choose which market to invest, and then to make investment decisions. All participants who choose market B will also make monitoring decisions. Note that maximum number of tokens free of charge and monitoring fine per each extra token invested change with a number of people in a MARKET B, however monitoring cost stays the same. See the attached Table B with the fine and monitoring cost in MARKET B. EXAMPLE #3 Suppose you (ID#1) and two other people invest in MARKET A with a total group investment of 200 and seven other persons decide to invest in MARKET B with a total group investment of 200 tokens. ID#
1
2
3
Marker A Market B
40
100
60
4
5
6
7
8
9
10
40
40
25
26
24
22
23
Total group investment 200 200
Therefore, return per token invested is the same, 7.83 francs, in both MARKETS since the total group investment is identical. Suppose ID#5 inspects ID#4, therefore receives monitoring revenue, 82.24=(40-24)x5.14 (see Table B for fines and maximum fine-free number of tokens). At the same time ID #5 pays a fine for extra 16 tokens=40-24, because someone else inspects her as well. After the market period everyone will know that ID #5 and ID #4 have invested 16 extra tokens each. ARE THERE ANY QUESTIONS?
Table B. Monitoring fine and cost in MARKET B Number of people in a Market 1 2 3 4 5 6 7 8 9 10
Fine-free Maximum number of token per person 0 87 58 43 34 28 24 21 18 16
Fine per extra token
Monitoring cost per person
0.00 3.00 4.00 4.50 4.80 5.00 5.14 5.25 5.33 5.40
0 7 7 7 7 7 7 7 7 7
6
INSTRUCTIONS (VV) PART 4:
instructions for the two-market voting condition VOTING
In this part of the experiment, you will first decide which MARKET to choose, A or B. After market choices, you will be asked to select, or vote for, exactly one of two alternatives, NO MONITORING and MONITORING. All participants in your market will make their decisions at the same time, without knowing the choices of others. The alternative that collects most of the votes will be the outcome in your market. If there is a tie between two alternatives, both alternatives get the same number of votes, and then the outcome will be chosen randomly. The decisions made in the other MARKET have no effect on your earnings. EXAMPLE #1 Suppose the number of participants in your chosen market is 6 and the number of votes for NO MONITORING is 4 and number of votes for MONITORING is 2, then NO MONITORING will be the outcome. EXAMPLE #2 Suppose the number of participants in your chosen market is 6 and the number of votes for NO MONITORING is 3 and number of votes for MONITORING is 3. This is a tie. Then the outcome will be chosen randomly between NO MONITORING and MONITORING. Please, go through the review question and fill in the blank lines with the values you think are correct. PRACTICE ID ______ REVIEW Consider the following voting decisions: Suppose MARKET A has 4 people and MARKET B has 6 persons. Given the numbers of alternatives listed below find the outcome of voting in each market. Number of votes for NO MONITORING in MARKET A: 3 Number of votes for MONITORING in MARKET A: 1 The outcome in market A:____________________________ Number of votes for NO MONITORING in MARKET B: 3 Number of votes for MONITORING in MARKET B: 3 The outcome in market B: ____________________________ Please, raise your hand if you have any questions and an instructor will assist you.
ARE THERE ANY QUESTIONS?
7
![free [download] biomechanics of sport and exercise with web resource ...](https://p.pdfkul.com/img/300x300/free-download-biomechanics-of-sport-and-exercise-w_5a2e06b61723dd9f2dc65799.jpg)
