1 Introduction

T

Owing to repeated interactions, the members are able to get better information about the activities and intentions of other members in the community than outsiders. This then results in better collective action or community participation through peer monitoring and social sanctions. In other words, the presence of social capital eliminates the transaction cost associated with monitoring and enforcing formal contracts. Moreover, it is now widely held that efficient functioning of formal institutions is also dependent on the degree of social capital. It is not difficult to find examples of identical institutions functioning efficiently in one region, while plagued with corruption and inefficiency in other regions, due to the adequate presence of social capital in the first case and inadequate presence in the other.

Members of poor communities in the less developed world

o ften critically depend on such social capital for their survival. Such social capital may take the form of social informational networks or collective efforts to maintain and regenerate common infrastructures or environmental services. In order to concretise the point, we introduce here the concept of a community good. This is the same as a public good (with consumption being nonrival and non-excludable) with two important differences. First, the good is non-excludable to all members of a community but persons from outside the community can be effectively excluded. Second, unlike a public good, which cannot be provided by the market but is usually provided by the government, a community good cannot be provided either by the market or by the government. This is because of their scattered and diverse nature. The only way in which such a good can be provided is if members of a community contribute time, money and resources to create the good. Here a community is defined by some socio-cultural marker and the logic of the emergence of such a community is outside the scope of enquiry of this paper. Examples of such community

Economic & Political Weekly

EPW

SPECIAL ARTICLE

goods may be microfinance groups, providing neighbourhood night patrol, management of a local watershed or a protection programme to regenerate a local forest. The question asked here is under what conditions will or will not members of a community come forward to create such a community good.

The adoption of microfinance schemes in several developing countries and promotion of self-help groups (SHGs), especially under the Swarnajayanti Gram Swarojgar Yojana (SGSY), in rural India to alleviate poverty, testify to the growing importance of development through group formation and collective action. The SGSY, first introduced in India in April 1999, is a clear shift from the individual approach to a group approach. The primary target of this scheme is to raise the income of those below the poverty line by providing them with resources and skills, and utilising the social capital of the community.

Moreover, people living in many developing countries are heavily dependent on agricultural and pastoral activities. These, in turn, are dependent on common property (pool) resources (CPRs) like water for irrigation or common land for grazing. The CPRs are also a source of subsistence for the asset-poor individuals and communities. They provide a safety net for the poor and play an important role in poverty alleviation. Empirical studies conducted by Jodha (1986) and Cavendish (2000), show that 15 to 20% of the income of poor families in the arid districts of rural India and 35 to 40% of income in rural Zimbabwe, respectively, are directly derived from the CPRs. But since these resources are commonly owned and used, there is little private incentive to maintain or regenerate them. In the absence of collective action or community participation towards their maintenance and r egeneration, the resources are likely to be degraded, further reducing the income base of the poorer section of the population. Along with it, other environmental and ecological services

o btained from these resources would also get restricted. This makes the maintenance and management of CPRs a matter of great concern for policymakers working for poverty alleviation. In particular, management options must include prevention of degradation of commons through local uses. Further, such policies must also address how commons already degraded may be regenerated. In this community management of CPRs provides an option. Here it may not be out of place to recall the success of the efforts initiated by the government of West Bengal to regenerate and protect forests by involving local communities in the Joint Forest Management scheme.

2 Some Problems of Community Participation

The difficulties of energising community participation or collective action in general, and in managing CPRs in particular, have been brought out in three important models. The first is that by Hardin (1968), in which he discusses why it is not possible to maintain and regenerate CPRs through peoples’ participation and concludes, “Ruin is the destination towards which all men rush, each pursuing his own best interest in a society that believes in the freedom of the commons”. The second model (Tucker 1950) that comes to a similar conclusion is the non-cooperative choice leading to a prisoner’s dilemma in which the dominant strategy for each player is to defect and so overall defection is the ultimate

52 outcome, though cooperation would have improved the condition of both players. The third important model is that developed by Olson (1965) in his well-known The Logic of Collective Action. Here Olson challenged the optimism expressed in group theory that individuals with common interest would act together t owards achieving their common target. According to Olson, if individuals cannot be excluded from enjoying the benefits of a collective good, once the good is produced, they have little incentive to contribute voluntarily to the provision of such goods. O lson further held that if the number of participants is small enough to be monitored by each other or there is some form of coercion to force all to contribute, cooperation might be possible.

The central assumption in the above three models is that individuals behave in a self-interested manner, which prompt them to free-ride on others’ effort. There is another type of game called the coordination game, where cooperative outcomes are not reached because of coordination failure among the players. Despite the problems of free-riding and coordination failure, there is a growing evidence in the efficacy of community participation which emphasises that in many cases communities may prove to be useful where both markets and state may fail. Bowles (2005) states that in many situations communities can effectively and costlessly monitor the behaviour of its members since they have more information about its members than outside agencies such as banks or the government. Thus, the communities are in a better position to take care of problems of strategic defaults in the credit market or problems of free-riding among its members.

3 Empirical and Experimental Evidence

In fact, there is much empirical evidence (Wade 1987, Baviskar 1994, D’Silva and Nagnath 2002, Moose 2003), which shows that village communities have been able to develop stable rules to stop degradation of CPRs and even regenerate them. For instance, D’Silva and Nagnath describe how the villagers of Behroonguda in Andhra Pradesh, India, resolved to protect and regenerate the forest around their village. They provided night patrols in groups to check encroachment of forests and stop timber theft. Also there are examples of more than 17 million rural poor who have formed SHGs and in many such groups, especially in Andhra Pradesh, savings of individual members have crossed Rs 10,000.

Besides empirical evidence of community participation, there is a host of experimental evidences to show that community participation may be forthcoming under certain institutional, social and historic situations. But before we get to such examples we provide a brief description of experimental economics.

Experimental economics can be broadly defined as testing in a controlled setting, well-formulated hypothesis concerning behavioural assumptions of economics (Bowles 2005). Laboratory experiments, according to Smith (2003), are methods of investigation into the “human interactive decision behaviour in social context governed by explicit and implicit rules”. The experimenter controls the explicit rules by framing the rules of the e xperimental games or through specific information supplied to the subjects. The implicit rules like norms, tradition and habits are beyond the control of the experimenter and the subjects bring them to the lab. Traditionally economists have been passive

may 16, 2009 vol xliv no 20

SPECIAL ARTICLE

o bservers of human economic behaviour and have derived results through deductions from axioms arrived at introspectively. Experimental economics provides a scope for verification of some of the theses and is an important source of data generator through controlled experiment.

Results of experimental games like the Ultimatum Game, D ictator Game, Trust Game, Public Goods Game and so on show that there exists an inconsistency between the theoretical conclusions of complete information Nash equilibrium predictions and actual observed results. Such apparently “irrational” behaviour has been explained in terms of social preferences exhibited by the agents. In other words, individuals do not only care about their own gains, but also the gains derived by others. These social preferences may be derived from alternative assumptions about the motivation behind individual behaviour such as fairness, a ltruism, reciprocity and so on. Moreover, there is a limit to human ability to engage in complex exchanges. In such cases, decisions may be taken on the basis of a well-accepted social norm or past experiences in similar situations or, sometimes, on the current state of the economic agent.

The prediction of overall defection or complete free-riding is based on the assumptions that one’s gains are the sole determinant of utility and that the individuals maximise utility functions which are monotonically increasing in experimental earnings. But as we have noted earlier, this may not be true under all conditions. There may be many unseen and unconceived processes, which may have affected a particular social phenomenon. For instance, Rabin (1993) and Levine (1997) have incorporated reciprocity or fairness considerations in shaping the preferences of individuals. Palfrey and Rosenthal (1994) model “uncontrolled” preferences where utility is derived from acts of altruism through social cooperation and contribution.

4 Experiments on Community Participation

The objective of this paper is to verify the logic of community participation or collective action under varying experimental conditions using a public goods game. We use the Voluntary Contribution Mechanism (VCM) public goods game. The experiment was one-shot, non-computerised lab games. The details of the e xperimental design are given in Section 6. The public goods game has been widely used to capture the tension between private and social gains in social dilemmas. A typical public goods game r equires the participants to contribute a part of their e ndowment (provided by the experimenter at the beginning of the game) to a common fund which is then multiplied by a factor (say doubled) and equally distributed among the participants. The participants can contribute any amount or take home the e ndowment. Thus those who do not contribute keep their initial endowment and also gain an equal share from the common kitty. Thus, for a self-regarding individual the dominant strategy would be to contribute nothing and by symmetry, the total contribution should be zero.

Ledyard (1995) in his survey of public goods experiments with VCM concludes that some of the major findings are that in both one-shot trials and initial stages of finitely repeated trials, subjects contribute halfway between the Pareto-efficient level and

Economic & Political Weekly

EPW

the free-riding level. But contributions decline with repetition. Similar observations have been made by Camerer and Fehr (2003) that lab experiments on public goods games in developed countries have shown that in one-shot games the group contribution ranges between 40% and 60% of the total endowment. A ccording to them, this suggests that subjects do not exhibit selfinterest in general, but behave reciprocally. This is further substantiated in repeated public goods game, where the group distribution d eclines in successive rounds. This is because when subjects who have made high contributions find that others have e ither not contributed or contributed low, they start reciprocating in subsequent rounds and group contribution declines.

5 Laboratory Experiments

The importance of community participation, reflecting the presence of social capital and its role in community governance has already been highlighted in the introductory section. We have conducted three sessions of public goods experiment to verify community participation among people in a developing country. As we have already discussed in Section 3, according to Smith (2003), the experimenters can only control the explicit conditions by framing different treatments under which the experiment is conducted. But the implicit conditions are beyond the control of the experimenters and are brought into the laboratory by the subjects. This implies that the outcomes of the experiment are subject to two sources of variation, that is, the explicit and the implicit conditions. To capture the impact of the explicit conditions there were four different treatments, which were played for one period each (thus constituting four one-shot games). Moreover, the experiment was conducted in three separate sessions drawing subjects from different socio-cultural backgrounds.

The first session of the experiment was conducted at Jnan Chandra Ghosh Polytechnic (JCGP) in Kolkata on 3 December 2005 with 30, first-semester diploma, students of electronics and telecommunication engineering, civil engineering, electrical engineering and mechanical engineering. There were four girls in this group. The age group of the students was between 17 and 19 years with average monthly family income less than Rs 6,500. The students were either from Kolkata or suburbs near Kolkata.

The second session of the experiment was conducted on 27 December 2005 with residents of Sundarika village, Diamond Harbour II, South 24 Parganas, West Bengal. The gender composition was six female and 24 male participants. The participants were members of a community-based organisation (CBO), a local club, with some history of community actions like social forestry, environmental awareness campaigns and voluntary blood donation camps. The average monthly family income of the participants of this session was less than Rs 3,000. The last session was conducted on 16 March 2006 at Phulbari in Sagar Island, South 24 Parganas. In this session the participants were five members from six different CBOs located in six different villages. Among the subjects there were four female participants and the average monthly family income of the participants in this session also was less than Rs 3,000. The subjects have participated in different

SPECIAL ARTICLE

s ocial and cultural projects undertaken by their respective CBOs, but have not participated jointly in any earlier project.

6 Design of the Experiment

The games in the above three sessions were one-shot VCM games under varying game conditions. That is, only one round of game was played for each treatment. The design of the experiment is similar to the Simple Public Goods Experiment described by L edyard (1995).

The subjects voluntarily participated in the experiment in response to an appeal circulated in the classrooms of the students and the same was affixed in the notice boards of the CBOs. They were selected on the first come first selected basis. On the days of the experiment they were requested to assemble in a classroom and the rules of the games were explained to the participants in Bengali and Hindi before the beginning of the game. The first author read out the same instruction sheet, containing the rules of the experiment, before the beginning of each session (given in Appendix, p 57). This was followed by a brief question-answer session to clarify any doubts about the games. However, the changes in the conditions of the game were informed before the commencement of each treatment. At the villages the experimental sessions were carried out in local high schools. In all the sessions each participant was provided with an endowment of Rs 50 at the b eginning of each treatment and asked to contribute from it t owards a group fund, which was to be doubled by the experimenters, and distributed equally among the participants. The group funds in the VCM games represent the public goods and the return from obtaining the public goods is captured by multiplying the kitty by a factor greater than one. Since public goods (or community goods in our case) are non-excludable in consumption, the magnified fund was distributed equally among the p articipants, irrespective of their contribution.

The endowment distributed was money (Rs 50) in unmarked and identical envelopes and each participant was to contribute an amount (say, Rs C, 0 ≤ C ≤ 50) towards a group fund. The d enomination of the endowment was four notes of Rs 10, one note (or coin) of Rs 5, two notes (or coins) of Rs 2 and one coin of Re 1. Thus, the participants could contribute any amount they wanted. It was clearly explained that the endowment supplied belonged to them and contribution to the common fund was

o ptional. They were to make their contribution behind a screen and the contributions were made in the envelopes and dropped in a basket that was also kept behind the screen. Thus, the contributions to the group funds were made without identification and the results of different games were withheld till the completion of the entire experiment to prevent results of earlier rounds influencing later rounds. Collecting the contributions in separate e nvelopes allowed the experimenters to obtain the distribution of contributions in each game.

As we have observed in the preceding section, in repeated games with VCM it has been generally found that the subjects contribute more generously in the earlier rounds but the average contributions decline in the later rounds. This has been explained as reciprocity on the part of the earlier contributors when they realise (after knowing the results of each period) that others have not been as

54 generous. Since our experiment was one-shot games with different treatments it was important for us to prevent the reciprocity factor from influencing the results of the later games so that the impact of the varying treatments can be isolated. Thus, we withheld the results of the different games till the end of the session.

Finally, the participants of the three sessions were requested to fill up a survey sheet in which they were to give the name of their village, monthly family income range, caste, religion, which p olitical party (if any) they considered to be best, whether they were involved in cultural activities and social activities. However their name was optional.

7 Hypotheses Tested

Before we present the hypotheses a comment about the lack of iterative repetition is in order. As we have mentioned earlier, the design of the experiment follows the Simple Public Goods Experiment described by Ledyard (1995), which could be run either one-shot or repetitively. Moreover, Ledyard (1995) and Camerer and Fehr (2003) observe that the outcomes of the earlier rounds of the repetitive games were close to those in one-shot games. However, more and more experimenters are using the repetitive method in the same treatment to generate more data and to increase the robustness of the estimates. In the present case this was not possible for two reasons. First, the experiments were non-computerised and the participants were required to physically go behind a screen and make the contribution one by one. Therefore each treatment required over 45 minutes and the whole session (including reading out of initial instructions, final calculations of contributions and making payments to the participants) about five hours. So there was not enough time for repetition. Second, experiments carried out over long periods of time cause mental fatigue to participants that lead to failure in cognition at later stages of the session. Hence, repetition is not advised in non-computerised experiments.

To find the impact of varying group size, the VCM was repeated thrice with group sizes 13, 21 and 30, respectively. This peculiar group break-up was necessitated by the fact that there were 13 students from electronics and telecommunication, eight students from civil and nine students from electrical and mechanical engineering in the first session that volunteered as participants. For comparability the same group sizes were maintained in the second and third sessions. These games were coded as 1a, 1b and 1c, respectively, in Table 1. All the remaining games had a group-size of 30.

Table 1: Game Codes and Brief Description of Games

EPW

SPECIAL ARTICLE

The next game was VCM with identification but without any punishment. Here the hypothesis was to examine the impact of peer criticism or sense of embarrassment on being revealed a free rider. This game has been coded as Game-2. As the conditions of the second game required identification of the participants so the endowments were distributed in numbered envelopes. The subjects were given envelopes bearing the same number as the serial number against their respective signature on the attendance sheet. They were informed that their names and contributions for the game would be put on the board at the end of the session. However, it was made clear that there would not be any kind of punishment or penalty. Moreover, no pre-play communication among the participants was allowed except before the beginning of the third game.

The third game was to find the impact of pre-play communication on total contribution. This was achieved by announcing a tea break of 20 minutes where subjects were asked to discuss their experience from the earlier rounds in the absence of the investigators. Finally, in the last rounds of the game, the students in session 1 were asked to contribute for cricket kits to be donated to the Students’ Union and the participants in session 2 were asked to donate for an ongoing social forestry project in their village. The participants in session 3 were to contribute for game kits for their respective CBOs. That is, the total contribution after doubling was divided into six parts. This game (Game-4) was conducted to test the impact of contributing for a real project. A brief description of the games with game codes is given in Table 1.

8 Results

The group contributions as a percentage of the total endowment for each game are presented in Table 2. Considering Game-1c as the game of reference, the following points may be noted; (i) under no circumstances was the contribution below 50% of the total endowment; (ii) in session 1, with students, the percentage of group contribution declined with increasing group size but no such trend was observed in session 2 or session 3; (iii) removal of anonymity and allowing pre-play communication increases group contributions in all the sessions; (iv) the contribution of the participants for the real life project declined in session 1 but increased in the other two sessions; and (v) even though the participants in session 2 and session 3 have almost identical sociocultural and economic backgrounds, the participants in session 2 have contributed higher in all the games.

Table 2: Percentage of Total Endowment Table 2 presents the Contributed to Group Fund in Lab Experiment

general findings from the

Game Codes Session 1 Session 2 Session 3

experiment. But before

(at JCGP) (at Sundarika) (at Sagar Island)

ground of the participants in the different sessions. The participants of the first session were first semester students who have been studying in the same institution for only six months and were from different departments and they would be staying in this institution for the next two and a half years. But the participants in the second session were

Economic & Political Weekly

EPW

from the same locality and knew each other and interacted with each other regularly for a long time and would be staying together (probably) for the rest of their lives. Moreover, being members of CBO and being from the same locality, they have experience of working together and a history of sharing and using CPRs. The participants in the third session had more or less the same characteristics as those of the participants in the second session as they too were members of CBOs, which undertook different social projects through member participation and a history of sharing and using CPRs in their respective villages. But, here they had to participate with members of other CBOs from other villages with whom they had little or no previous interactions or would have little in the future.

Now it would be easier to interpret the results of different rounds of Game-1 in the three sessions. From Game-1 no clear picture emerges. Even though the group contribution declined with increasing group-size in session 1, no clear trend was observed in the other two sessions. But the results of Game-2 reveal that there is a positive impact of identification (even without any explicit punishment) on group contribution. This may have r esulted from tacit threat of peer criticism or from a sense of e mbarrassment at being revealed a free-rider. In the third game also, we find a positive impact of pre-play communication in all the sessions. In the third game, communication among the participants provided individuals better information about others’ participation and as a result their perception about others’ participation has increased. Finally, we find that when contributing for a real project (the returns of which will not be immediately obtained) the villagers show more cooperation than the students do.

However, the degree of response between the participants in session 2 and session 3 is starkly different considering the fact that the participants have more or less similar socio-cultural backgrounds. The group contributions by participants in session 2 have been higher in all the treatments than those by participants in session 3. This may be attributed to a common bondage with the locality and the community among the participants in the session 2 as they come from the same CBO and have a history of working together in earlier social and cultural projects. But the participants in session 3 were from six different CBOs without any history of working t ogether, consequently, the bondage between the members of the dif-

Table 3: Result of Analysis of Variance

though the participants in the last two sessions came from more or less similar socio-cultural backgrounds the societal structures of the two groups were different.

The contributions as percentage of total endowment presented in Table 2, are outcomes of two factors, the different treatments and the different social and cultural structure of the groups. To find out the effects of the two factors we carried out a two-way ANOVA and the result of it is presented in Table 3.

is 3.33, while F0.05, 2, 10 is 4.10 and F0.01, 2, 10 is 7.56. The

F0.05, 5, 10 observed F for difference in treatment being smaller than F0.05, 5, 10 is statistically insignificant at 5% level. But the observed F for

SPECIAL ARTICLE

d ifference in societal structure is greater than F0.01, 2, 10 is statistically significant at 1% level. This shows that the impact of different treatments on mean percentage contribution is identical in all three sessions while it varies with socio-cultural structure of the groups. Additionally we have also carried out paired t-test to examine the equality of means in the different sessions. The results are shown in Table 4.

The high pro-social behaviour exhibited by the subjects in session 2 seems consistent with findings in a threshold public goods game at Peru, as reported by Cardenas and Carpenter (2005). At least 81% of the participants contributed and the value of contribution was naturally high. Here it may be also interesting to note that Cardenas and Carpenter refer to experiments which show

that in prisoner’s dilemma games Chinese stu-

0.05, 5

imply that the difference of socio-cultural structures of the groups in sessions 1 and 2 and sessions 2 and 3 affects average percentage contribution.

9 Concluding Remarks

We end this paper by making a couple of brief observations. First, it should be noted that the group-contributions in all the sessions were above 50% and those in session 2 were always higher than in session 1. Second, even though the participants in sessions 2 and 3 have more or less similar socio-cultural and economic background, their group contributions were significantly different. In case of the villagers, in session 2, the contributions have been above 80%. But as reported in the literature on public goods game, the average contributions in developed countries lie b etween 40% and 60%. Here we try to explain and also justify these differences in contribution between students in an urban technological institution and members of a village CBOs in terms of differences in societal structure.

Grief (1994) shows that institutional structures and intersociety adoption of institutions are dependent on culture. The collective societies are characterised by a segregated social structure, in which social and economic interaction takes place mainly among the members of that specific society. Cultural beliefs then become identical and common among the frequently interacting members. The members of these societies feel involved in the lives of fellow members. This kind of closed social network within a homogeneous community helps create social capital. The individualist societies, on the other hand, are characterised by integrated social network. The individuals have social and economic interactions with members from different groups and may shift from one group to another. Members of such heterogeneous communities have lesser attachments for members of any specific group and self-reliance is highly valued in the society.

than in Thailand. This, according to them, may have been because of the collectivist culture advanced by the governments of China and Vietnam compared to individualism promoted in United States and Thailand.

This high pro-social behaviour stands in sharp contrast to that exhibited by the students in session 1. As we have stated earlier, the students were fresher and were from different streams of engineering. As a result, they had little scope of interacting among themselves. Moreover, being city dwellers and exposed to western system of education they may have been influenced by individualistic societal structure. On the other hand, the participants of session 2 were from the same CBO. Repeated interactions and a history of successful cooperation may have helped them to have better perception about others cooperative behaviour, which, in turn, may have resulted in high group contributions. That is, they behaved like a closed homogeneous community.

The difference between the group contributions in sessions 2 and 3 may be explained by what Grief (1994) has termed as a segre gated social structure in developing countries. In these situations there exists a strong social bonding within members of a community or group but not with outsiders. Dasgupta (2005a) makes a similar observation. According to him, social capital is a private good with both positive and negative externalities. Positive externalities can spread beyond immediate groups but negative externalities, like community cohesion developed on the basis of caste, religion or race can take place at the cost of outsiders. In the latter case, outsiders may even be treated with hostility and suspicion. As the subjects in session 3 had to participate with members of other CBOs with whom they had little previous interaction, they had very little information about the members from the other CBOs. They may have behaved as segregated communities which has reduced the social cohesion or bonding within the group.

Baviskar, A (1994): “Fate of the Forest: Conservation and Tribal Rights”, EPW, Vol XXIX, No 39, 17 Sep-

Bowles, Samuel (2005): Microeconomics, Behavior,

Camerer, F Colin and Fehr, Ernst (2003): “Measuring Development, 28(ii), pp 1979-2003. S cience, 162, pp 124-42.

coming in Henrich, Boyd, Bowles, Camerer, Fehr,

may 16, 2009 vol xliv no 20

SPECIAL ARTICLE

prepared for the WIDER Jubilee Conference

(Thinking Ahead: The Future of Development

Economics, Helsinki, 17-18 June 2005). Ledyard, John O (1995): “Publics Goods: A Survey of

Experimental Research” in Kagel and Roth (ed.),

The Handbook of Experimental Economics (Princ

eton University Press). Levine, D K (1997): “Modelling Altruism and Spiteful

ness in Experiments”, Review of Economic Dyna

mics, Academic Press for Society of Economic

Dynamics, Vol 1(3), pp 593-622.

Appendix Instruction Sheet

(translated into English from Bengali)

Welcome to a session of Economic Experiment on Community Participation. As all of you know that there are many things which cannot be obtained through individual effort alone and require participation from other members of the society. Examples of such things are not difficult to find. If we take the example of planting trees for a social forestry project, it cannot be undertaken by an individual but requires the participation of other members in the community. Some may provide money, some labour and others resources. However, it is also observed that all members do not show the same interest in participating for these common goods. They may wait for others to obtain the goods. But since these goods are common goods, those who do not participate or contribute cannot be excluded from enjoying the benefits of these goods once they are obtained. That is, after the social forest is obtained, no one in the community can be prevented from enjoying the greenery or fresh air.

We have tried to capture this scenario of community participation through some experimental games. Each one of you will be provided with an envelope containing Rs 50 (four notes of rupees ten, one note (or coin) of rupees five, two notes (or coins) of rupees two and one coin of rupee one. You will be required to contribute some money (any amount between zero and Rs 50) from the money given to you towards a common fund. This money belongs to you and contribution is optional. However, the total group money collected will be doubled by the experimenters and divided equally among the participants (irrespective of their contribution). Thus, your earnings consist of the amount of money that you have not contributed plus an equal share of the doubled group contribution. We shall illustrate the calculation with an example on the black board.

The envelopes are unmarked and identical. You will have to make your contribution behind the screen. There you will find a basket. You will have to make your contribution in the envelope given to you and place the envelope in an upright position in the basket. Even if you do not make any contribution you must leave the empty envelope in the basket. You are requested not to discuss anything with other participants. In case you need any clarification please feel free to ask us.

Delhi: OUP).

Smith, V (2003): “What Is Experimental Economics”,

web site of Interdisciplinary Centre for Economic

Science at George Mason University.

Staub, Ervin (1978): Positive Social Behavior and

Morality (Academic Press). Tucker, A W (1950): UMAPS J, I, p 101.

Wade, R (1987): “The Management of Common Property Resources: Collective Action as an Alternative to Privatisation or State Regulation”, Cambridge Journal of Economics, 11, pp 95-106.

have to make contributions from it to a common fund. The group contribution will again be doubled and divided equally among the participants. At the end of the session your contributions in this round will be displayed on the board against your names.

Instruction Given before Treatment 3: Now we shall take a tea break for 15 minutes. As tea is being served we would expect you to discuss about the games and share your thoughts with other participants.

After 15 minutes the participants will be called back to the classroom.

Now we shall repeat the game of the first round. Please check that you have Rs 50 in your envelopes and that your envelopes are unmarked.

Instruction Given before Treatment 4: In this game you will have to again contribute to a group fund from the money inside the new envelopes given to you. The rules of the game remain the same except for the fact that the group contribution after doubling will not be divided equally among the participants but will be donated on your behalf to a community project.