Conditional Cooperation in the Field: Cross-Country Skiers Behavior in Sweden*

Please download to get full document.

View again

of 26
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information Report
Category:

Concepts & Trends

Published:

Views: 0 | Pages: 26

Extension: PDF | Download: 0

Share
Related documents
Description
Conditional Cooperation in the Field: Cross-Country Skiers Behavior in Sweden* Tobias Heldt June 3, 2005 Abstract In a laboratory one-shot public good game, Fischbacher, Gächter and Fehr (2001) classify
Transcript
Conditional Cooperation in the Field: Cross-Country Skiers Behavior in Sweden* Tobias Heldt June 3, 2005 Abstract In a laboratory one-shot public good game, Fischbacher, Gächter and Fehr (2001) classify 50 percent of the subjects as conditional cooperators. Outside the lab, using a student sample, Frey and Meier (2005) find that people behave pro-socially, conditional on others behavior. This paper tests for conditional cooperation and social comparisons in a natural field experiment, using decisions from a sample of cross-country skiers in Sweden on the issue of voluntary cash contributions to the preparation of ski tracks. Two test procedures are used. First, testing for correlation between beliefs about the contribution of others and own behavior and second, experimentally varying the beliefs about others behavior. Using the latter approach, I find the share of subjects giving a contribution to be significantly greater in the group receiving information about others behavior than in the group that does not. Regression analysis cannot reject that subjects are affected by social comparisons and express a behavior classified as conditional cooperation. JEL: H41, C93, Z13 Keywords: conditional cooperation, natural field experiment, public good, voluntary contributions Contact info: Department of Economics and Society, Dalarna University, Sweden *Acknowledgements: I thank Håkan Holm, Lars Hultkrantz, Olof Johansson-Stenman, Chuan-Zhong Li, Jan-Eric Nilsson and Maria Vredin-Johansson for useful suggestions and comments on the manuscript. I also thank Ernst Fehr and participants in the Summer school of Behavioral Economics, Sannäs June 2004, for useful comments on the presentation of the experimental design. Seminar participants at VTI/Dalarna University and Örebro University also provided useful comments. Thanks to Christer Rosen, Säfsen Resort, for making the field experiment possible. All remaining errors are my own. 1 Conditional cooperation is intuitive (Armin Falk, University of Zürich, 2004) we might all of us be willing to contribute to the relief of poverty, provided everyone else did. We might not be willing to contribute the same amount without such assurance. (Milton Friedman (1962) Capitalism and Freedom, p. 191) 1. Introduction Today, experimental evidence as well as real world empirical evidence firmly establishes that the self-interest prediction of zero contribution in a one-shot public good situation can be rejected (e.g. Ostrom 1990, Ledyard 1995). It is suggested that the existence of conditional cooperators is the source of initially high contribution rates in such public good games (Ostrom 2000). A theory of conditional cooperation states that higher cooperation or contribution rates are expected when information is provided that many others cooperate (Frey and Meier 2005). The theory suggests that people compare themselves with the behavior of a reference agent, i.e. they make social comparisons. That people would be willing to contribute to a public good conditional on other people s cooperative behavior, was found in Fischbacher, Gächter and Fehr (2001). In their oneshot public good game, 50 percent of the subjects were classified as conditional cooperators. Several laboratory experiments have identified behavior that has been categorized as conditional cooperation (e.g. Keser and van Winden 2000), but only Fischbacher et al. (2001) explicitly test for conditional cooperation. Evidence of conditional cooperation has also been found outside the laboratory. In a field experiment on charitable giving, using a student sample, Frey and Meier (2005) find that contributions increase, on average, if people know that many others also contribute. They test conditional cooperation in two ways. The first is by varying the beliefs about the behavior of the group in an experimentally controlled way, the second is to test whether expectations about the behavior of the group varies positively with one s own behavior. In this paper, I apply both ways of testing conditional cooperation in a natural field experiment, using decisions from a non-student sample. 2 The context of the experiment is a Swedish ski resort where a sample of cross-country skiers (hereafter skiers ) face a decision to voluntarily contribute to funding the grooming of ski tracks. Ski tracks in Sweden are public goods, since for undeveloped land, including trail infrastructures such as a ski track, it is not legal to exclude skiers. Moreover, it is not legal to charge for access to such facilities. In this study, one part of the skier population is defined as tourists, making a one-shot (one week) visit to a destination. The rest are local residents using the tracks for more or less regular training or recreational purposes. Consequently, I study a real life situation where voluntary contributions constitute a direct way of funding a public good, with a sub sample of the population playing a one-shot game. I get results that cannot reject the existence of social comparison and conditional cooperation among the group of subjects defined as tourist skiers. When experimentally varying the beliefs of others behavior, I find the share of subjects giving a contribution to be significantly greater (p 0.10 level) in a group receiving information about others behavior than in a group not receiving any such information. A series of regressions support the result. Testing conditional cooperation in a natural field experiment has at least two clear advantages as compared to previous studies. First, while it is still unclear how results from laboratory research can be generalized outside a laboratory situation, this field experiment adds to narrowing this gap. Second, as compared to studies using a student sample, for example Frey and Meier (2005), this field experiment can control for variability in socio-demographic characteristics. This is relevant since, for example, age has been shown to positively correlate with social preferences in general and contributions in a fundraising drive in particular (List 2004). Moreover, the Frey and Meier (2005) field experiment used a context of altruistic donations to charity, while the present field experiment is, to my knowledge, the first to test conditional cooperation in a context where subjects actually get to use the public good to which they contribute. 3 Finally, the study can be categorized as a natural field experiment following the definitions of Harrison and List (2004). The remainder of the paper is organized as follows. Section 2 presents the field experiment context and the data gathering process. Section 3 reviews related literature and presents hypotheses to be tested. Results and analyses are found in section 4. Section five concludes. 2. Field experiment context and collection of data The field experiment uses data of individual human behavior at Säfsen, a Swedish ski resort. Data were collected during eight weeks in January and February Arriving in the area with prepared tracks, each skier was given the opportunity to voluntarily contribute to ski-track funding. However, depending on their degree of attachment to the area, skiers might perceive the field situation differently. Three categories of skiers were expected; locals living close to the area and using the ski facilities regularly, semi-locals living rather close and making a few day trips to the area during a season for training purposes and tourists, visiting the area possibly only one week or weekend during a season. To motivate the different types of behavior, the first part of this section begins with a description of the field context and the expected behavior of different categories of skiers. The second part describes the experimental design. Cross-country skiing in a tourist area in Sweden is an example of the general problem of tourism that uses natural and cultural environments with limited possibilities of excluding visitors. In Sweden, public access to undeveloped landscapes is open under Allemansrätten, a centuries-old right of common access. It is not explicitly codified in law, but is rather a set of customary rules and judicial interpretations regarding activities, such as camping, berry picking, and cross-country skiing, not listed among state obligations to defend landowner claims. Facilities and infrastructure on land, for example snowmobile trails, mountain bike trails, canoeing waterways or cross-country trails then 4 get the features of public goods. 1 The free-rider problem stemming from open access is at the root of the problem of developing tourism that uses natural and cultural environments, mainly because it is difficult or impossible to price the actual use of the products or services. Investing in quality enhancing activities or facilities would give a product attracting more tourists, although it may be infeasible to recoup the costs invested. High-quality ski tracks are dependent on regular investments in track grooming in wintertime and some preparatory work in summertime. Rational self-interested people have little incentives to contribute to these investments. However, numerous case studies have shown that successful local governance of open access resources is possible (Ostrom 1990). Two criteria found to encourage cooperative behavior and result in successful governance are (i) a small number of members in the group and (ii) strong mutual obligations and ties among members (Grafton 2000 p.506). Tourists visiting a destination are typically not likely to meet any of the two criteria, while regular local and semi-local users (the sweat-shop bugs) are much closer. I assume that being a tourist 2 implies making a one-shot visit to a destination. It is well known that in a one-shot public good game, people have incentives to free ride. However, it is not unusual that tourist destinations have a high percentage of return visitors. But it is uncertain whether tourist visitation rates at a destination of once a year or once every second year change this incentive. Considering that tourists have many other possibilities to contribute to a destination s economy, for example by buying food, lodging etc., it seems reasonable to make the assumption that a tourist has the incentive to free ride when confronted with an option to voluntary contribute to a recreational public good in a tourist area. 1 Recreational land has been defined as an impure public good due to the aspect of rivalry in land use. Viewed as a common pool resource, recreational land is subject to reciprocal externalities in the form of short-term congestion and long-term depletion (Vail and Hultkrantz 2000). 2 A tourist is defined by the UN as a visitor who has traveled to and visits a destination different from her home environment. The length of the stay must not exceed one year and the purpose of the trip should be something else than to act as paid labor (World Tourism Organization 1994). 5 2.1 Experimental design, field conditions and data The experimental design has its origins in the unique feature of one Swedish ski resort where the norm of skiers contribution to fund ski-track grooming has been established. Heldt (2005) reports that 70 percent of the skiers at this particular ski resort contribute to track preparations. Since a previous study (Heldt and Nerhagen 2001) showed the limitations of eliciting varying degrees of group behavior using only one ski area, I use the information on 70 percent voluntary contribution from the unique area and post it as treatment information in my field experiment ski area. This was not a deception since that study and the one described in this paper ran in parallel. An important practical part of the field experiment was to present skiers with information about the group behavior before they decided to contribute. The field experiment included two periods: one control and one treatment period. In the control period, subjects received no explicit information about group behavior. In the treatment period, subjects were provided with information about the share of voluntary contributions within the group of skiers (the group contribution ), i.e. 70 percent. General information on the motivations for fundraising was present during both control and treatment periods. 3 To reduce the risk of data contamination, the control period had to come first and the treatment period second. Moreover, it was not possible to include a second explicit treatment with a variation in average group contributions. The reason is typical for a natural field experiment and is related to the credibility of the field context. Any rumor of manipulations or unseriousness would have discredited the study and would have had implications on subject behavior. In the control group, subjects beliefs about the share of voluntary contributions within the group were elicited. It is obvious that this information could not be gathered during the treatment period. 3 See Appendix 2 for the exact wording of general information and treatment information. 6 Field conditions The ski resort had no history of skiers contributing to fund ski-track preparations. Rather, the tourist resort finances ski-track preparations out of its general revenue. From this season, there was an option of voluntary contributions. When entering the track, skiers were informed about the motivations for the new voluntary contribution system. The information included an explicit request for a voluntary contribution of 20 SEK ($3 4 ) to fund track preparations. There were no explicit statements that the total amount contributed was in any way connected to the quality of present or future tracks. Cash contributions were to be made in a locked box, close to the information sign. Since both the subjects decision of whether to contribute as well as subject category were of interest, a questionnaire was used to gather relevant information. This also made it possible to collect other individual-specific data. Using a questionnaire, the potential problem of sample selection must be considered. There may be several reasons why a person is not responding to a questionnaire survey and some characteristics or behavior of the respondents may differ from those of the nonrespondents. Selection bias might be a problem in this study, if the subjects who did not contribute did not answer the questionnaire. To minimize the risk for selection bias, it was specifically pointed out that participation was voluntary and answers were anonymous. To promote a high participation and response rate, subjects were given a lottery ticket together with the questionnaire. Moreover, the stated purpose of the questionnaire being part of a research project in nature tourism at a university, has previously been shown to promote a high response rate as compared to marketing surveys with business purposes. All skiers at a specific location were sampled at randomly chosen time intervals. Subjects were asked to complete the questionnaire on site and return it in a box at a discrete location. The normal procedure was that a skier arrived by car, entered the ski track, passing closely by the information signs, took a ski trip, returned to the parking lot and 4 $ 3 is an approximate value at an exchange rate of 7 SEK/$. 7 was sampled for the survey. Reluctant participants could be provided with a pre-stamped envelope and be offered the alternative of completing the questionnaire at home and returning it by mail. To screen subjects that had been sampled multiple times and still keep the anonymity, the questionnaire asked subjects to state the last three digits of their car s license plate number. Quite often skiers go in groups. To enable the use of only independent observations in the analysis, questions of party size and age were included. To increase the participation rates, the questionnaire was deliberately made short. After several pretests, the final version had an average answering time of about five minutes. To capture subjects behavior, the question Before leaving this place, have you made a voluntary contribution of 20 SEK for track preparations? was used. To cross-validate reported with actual behavior, information from the ski resort on the total amount of money contributed in the box during each sampling period was used. The actual average contribution in seven out of fourteen sessions based on all sampled subjects was lower that the corresponding sum of reported individual contributions. However, by looking at the amount of money, it seems likely that a few subjects placed a lower contribution instead of the suggested 20 SEK. It could also be the case that a few subjects have reported a contribution but did not actually give one within the time of the sampling period. The contribution box was open to skier contributions 24 hours a day. The analysis in section 4 is based on reported behavior. A sensitivity analysis based on sessions with a perfect match between reported behavior and the corresponding sum of money in the box does not change the main results presented in section 4. During the treatment period of the field experiment, information stating that 70 percent of the skiers had been contributing was provided. To ensure that all subjects got the information, it appeared as a separate note on the information board as well as a separate note in the questionnaire. A total of 89 subjects were sampled for the treatment period, 38 of which entered the effective sample. Table 1 provides further details. 8 In the control period, no information was given on the group contribution. Instead, we included a separate question to elicit subjects own perception of the share of voluntary contributors within the population of skiers in the area. A total of 233 subjects were sampled in the control period, 89 of which entered the effective sample. The difference in sample size as well as the composition of skier categories between control and treatment is due to exceptionally good skiing conditions in the study area, which attracted an unusually large number of semi-local skiers. 5 Table 1: Effective sample sizes Control Treatment Total Responses a) Previously sampled excluded More than one in a party excluded Effective sample a) Response rate Control 76 percent, Treatment 75 percent. 3. Related literature and behavioral hypotheses Conditional cooperation is based on the notion that people are affected by, and compare themselves to, references groups, i.e. they make social comparisons. Conditional cooperators are people who are willing to contribute more to a public good the more others contribute (Fischbacher, Gächter and Fehr 2001 p. 397). A theory of conditional cooperation states that higher cooperation or contribution rates are expected when information is provided that many others cooperate (Frey and Meier 2005). Moreover, it has been suggested that a conditional cooperator does not only conditionally cooperate, but also initially cooperates when no cooperation norm has been established (Ostrom 2000). Ostrom s conditional cooperators are individuals willing to initiate cooperative actions when they believe that others will reciprocate. They are also willing to repeat 5 In the estimated models of section 4, I have also controlled for the potential influence of the Swedish sports break, without any significant effect. Treatment periods included three out of four sports break weeks. 9 these actions, as long as a relatively high proportion of other people involved in the situation reciprocate. Evidence of conditional cooperation can be found in several laboratory studies (Keser and Van Winden 2000, Fischbacher et al. 2001, Croson et al. 2004). However, not all people are expected to be conditional cooperators. Fischbacher et al. (2001) classified 50 percent of the subjects as conditional cooperators, while 30 percent were classified as free riders or purely selfish. Furthermore, Frey and Meier (2003) find that only ce
Recommended
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks