4 Cooperation

by Emma Hi’ilani Thain

Cooperation, in the context of the climate crisis, is understood within the game theoretic framework. Climate change is largely understood as a “collective action problem” or “collective risk social dilemma” (hurlstone, et. al, 2016). A collective action problem is a situation in which social welfare is at odds with individual self-interest. In order to achieve a socially optimal outcome, all individuals must cooperate to solve the collective issue, but if some participants don’t cooperate, the costs and benefits are unevenly distributed. Because there is always an incentive to act against the collective interest, collective action problems often end with a non-adaptive outcome for most or all participants. The famous example of this is the “prisoner’s dilemma,” in which two prisoners each benefit from outing the other. The most individually optimal choice for each is least optimal for the other, but the tendency is for each to choose the option that is least optimal for the other, which ends up being least optimal for both, while the socially optimal solution would have been okay for all. Cooperation is what is necessary for players to achieve the socially optimal outcome; it is when two or more individuals or groups work together to achieve a mutual goal in a way that maximizes reward and minimizes cost for all participants.

Climate change fits the format of a collective action problem, and indeed has been called “the ultimate global commons problem” because in the short-term, countries, businesses, and individuals benefit from continuing to live the same way that they currently do, which means emitting harmful gasses into the atmosphere and other unsustainable behaviors. A healthy climate is a global public good, making fossil fuel pollution a “tragedy of the commons” problem (hurlstone, et. al, 2016). “Tragedy of the commons” refers to a type of collective action problem in which there is a common good, such as a population of fish, or in this case, a temperate climate. The common good should be dealt with in a way that allows everyone to benefit from it, but each individual involved would benefit from taking more than their fair share. In the case of climate change, it’s more of a “common bad,” in which individuals must limit greenhouse gas emissions in order to preserve a temperate climate, but each is motivated to continue to emit because of economic incentives and a general convenience. However, our current behavior, if continued, will be catastrophic to the existence of our species and most other life forms on earth (Tingley & Tomz, 2014). If only some people or countries cooperate by taking action against climate change and changing their behavior, the situation functions negatively in one of two ways: either not enough people take action, rendering the actions of those who do take action as irrelevant, or enough people take action but some people, who did not cooperate but acted as “free riders,” receive the benefits of climate mitigation without incurring the costs. Both of these potential outcomes are theoretically deterrents for action.

When it comes to the climate crisis, “cooperation” is individuals, companies, institutions, and governments making extreme changes in policy and behavior to stop greenhouse gas emissions.

The vast majority of research on cooperation and climate change has been done in the context of cooperative “games.” These studies model varieties of game theoretic problems in a controlled setting, hoping to gain insight into the details of cooperation by mimicking real-world scenarios. These games generally feature either a common resource or a threshold. Common resource games are modeled after the “tragedy of the commons” problem, and require players to responsibly manage a common good. Threshold problems require players, over the course of some number of rounds, to contribute tokens, usually money, to a group account, with the intention of reaching a certain threshold. In each scenario, tension is created between individual interest and the socially optimal behavior, and different conditions are introduced to gauge if and why players cooperate.

The current work on cooperation and climate change originates both in psychological and economic literature, where cooperative games are often assessed, in political science, climate science, and sociology. The information in this chapter was drawn from many review papers, and thus represents some of these different fields. Most of these studies were done in the U.S., but some studies were done in a European or Chinese context.

A review of the literature reveals the following as significant factors that determine willingness to cooperate in collective action social dilemmas.

4.0.1 COMMUNICATION

An extremely reliable finding in the context of cooperative games is the benefit of communication between players (Sturm, et. al, 2019), (hurlstone, et. al, 2016). A study by Tavoni, et. al, demonstrates a poignant example. In a 10-round cooperative game, groups were supposed to collaborate to achieve a monetary contribution threshold. As a manipulation, participants were allowed to make non-binding pledges at the ends of rounds 3 and 7. These pledges declared their intent to give a certain amount of funds toward the collective goal and were shown to each consecutive player as they made their choices for the next round. In this open communication condition, 30% more groups reached the threshold than in the non-communicative condition. In another condition with heterogeneous participant resources, cooperation went up by 40% after communication was introduced (2011). This study is just one example of this highly replicable finding. Another example is illustrated by Brick, Visser, and Van der Hoven, who find that, on average, communication increases cooperation within a collective action dilemma. In their study, two dominant contribution norms emerge once communication is introduced: perfect cooperation and free-riding. With the ability to discuss contribution strategies, some participants consistently choose the socially optimal contribution. Alternatively, other participants, supposedly due to their confidence that other members of the group are handling the situation, begin to contribute less or not at all. Because of the high prevalence of free riding (26%-46% depending on personal cost of contribution), the authors underline the importance of sanctions for punishing non-cooperative players. Sanctions will be discussed further in a section below. Regardless, communication still significantly improved cooperation as a whole, with a 32% increase in the number of groups able to meet the necessary threshold compared with the non-communication condition (Brick, Visser, & Van der Hoven, 2015). Clearly, communication between participants is an essential piece of increasing cooperation. It’s hypothesized that communication increases other participants’ trust in their peers’ level of cooperation, subsequently improving their own willingness to act cooperatively (hurlstone, et. al, 2016). However, the climate crisis necessitates action from millions of people. It’s not feasible to imagine communication between everyone; it will likely be necessary to delineate groups that can actualize the benefit of communication. These groups may exist on a national level, such as countries agreeing on emission cuts, or on an individual level, such as members of a community agreeing to switch to solar power.

4.0.3 PLURALISTIC IGNORANCE

Pluralistic ignorance is a situation in which a person thinks that their privately held beliefs are contrary to the majority’s beliefs, when in actuality the majority also holds the same beliefs. The majority is individually afraid to openly oppose a norm because it appears that everyone else agrees with it, so the widely unsupported norm continues. Knowing what we know about communication and social trust, pluralistic ignorance can pose a serious obstacle to cooperation.

In their research, Geiger and Swim found a powerful expression of this phenomenon in a US university sample. In two studies, it was found that inaccurate perceptions of peers’ lack of concern for climate change caused concerned individuals to self-silence at a higher rate than those who perceived that their peers were also worried about climate. Those who thought that their peers cared about climate change were more willing to discuss it (Geiger & Swim, 2016). Pluralistic ignorance can be a hindrance to climate discussion and action, and reinforces norms of silence and the appearance of apathy. This poignant example once again highlights the importance of communication in a social dilemma, through which people can begin to accurately gauge their peers’ desire to cooperate.

In Lübke’s analysis of data in the EU, they found that cooperation beliefs in Europe are an underestimate of actual cooperation. On average, cooperation beliefs reflected the impression that fellow citizens contributed less to pro-environmental behaviors than they did in reality. This reveals what Lübke calls a “trust gap.” Generally, people underestimate how much their peers or co-citizens are doing to help mitigate energy use, based on self-report data of people’s behaviors. The participants in Lübke’s study are also displaying pluralistic ignorance, except in this case, the “privately held belief” also manifests in energy-saving behavior. The study does find that those with higher cooperation beliefs, which corresponds to a more accurate perception of citizens’ actual energy-saving behavior, are more likely to engage in self-reported energy-saving behavior. However, on average, they find that people’s energy-saving behavior supersedes their expectations about others’ cooperative action (Lübke, 2021).

4.0.4 PERCEPTION OF RISK

Simply put, perception of risk is the degree to which someone believes that a problem is serious enough that they should respond. Cooperation studies show that perception of risk is an important factor for participant contribution. Keeping in mind the abundance of misinformation and ambiguity that surrounds the climate crisis, this body of work signals the need to broadly clarify and communicate the real risk that the climate crisis poses to humanity.

In Milinski et. al’s seminal 2008 paper, they set out to test the influence of risk perception on cooperation levels in a threshold-style collective action game. In this study, they created three risk conditions. Participants faced either a 10%, 50%, or 90% risk of catastrophe if they failed to reach the threshold goal by the end of the game. The results indicated that the higher the perception of risk, the better. Under the lowest risk condition, no groups reached the threshold for catastrophe evasion. In the moderate risk condition, only 10% of groups reached the threshold. Finally, in the highest risk condition, 50% of groups reached the threshold and the rest came close (Milinski, et. al, 2008). The results of this study seem to indicate that it’s critical to make the extremity of climate risk as salient as possible, and that any downplaying can severely impact the vitality of the public’s response.

However, cooperation games are only a tool for predicting cooperation in the real world. It’s often a concern that communicating the real danger of the climate crisis will cause people to draw away from the issue altogether because the reality is too overwhelming to process. It’s possible that this study found extreme risk to facilitate cooperation because of three important considerations.

The participants in the study have very little ambiguity that it is their responsibility to respond to the risk, as they are participants in the study and it is literally their job to contribute to the solution.
The participants in this study have no ambiguity about which action to take and whether that action will work. These have been outlined by the nature of the study. They only need to contribute to the threshold and they will avoid catastrophe.
The participants are not facing a genuine risk. They are taking part in a study, which holds less stake, by orders of magnitude, than the real future of themselves, their species, and the planet.

Each of these three factors looks very different in our real collective action dilemma. It can be very unclear to any given individual or country that resolving the climate crisis is their responsibility or that there is an action that they can actually take that will actually have an effect on the issue. Additionally, it should be expected that emotional stress related to the climate crisis could be very severe. These factors should be taken into account when attempting to apply the results of this study to the real world. Communicating intense risk is likely only beneficial for cooperation if there is a somewhat clear and effective path to helping — otherwise it may only lead to panic.

4.0.5 EFFICACY BELIEFS

Efficacy belief is the assurance that individual and collective actions will actually be effective in solving the problem it’s intended to solve. Research shows that efficacy beliefs are a big factor in whether people actually engage in helpful behaviors (L. Steg, 2023).

4.0.7 REPUTATION

Research on cooperation games shows that social perception has a beneficial influence on cooperation. Essentially, when there is the potential that others will know about a player’s contributory action, the player is more likely to contribute more. This effect was found through the manipulation of a multiple-round threshold game, in which investments in the climate account were made public during 'odd' rounds and made anonymous during 'even' rounds. During the public rounds, players had an incentive to cooperate because they could improve their reputation. However, there was another incentive, the “reciprocity game.” This game occurred after the public round, and gave players the opportunity to give tokens to another player. Creating a good reputation by contributing in the game could incur the reward of a gift. During the anonymous rounds, the incentive to cooperate and the disincentive to free ride were eliminated, as reputation no longer played a role in gaining benefits in the indirect reciprocity game. Contributions were significantly higher in the public rounds than the anonymous rounds (hurlstone, et. al, 2016).

Reputation, as a mechanism for increasing cooperation towards climate mitigation, is a relatively easy tool to leverage, and has the added bonus of contributing towards a positive behavioral norm.

4.0.8 STRONG LEADERS

There is one study that suggests that the presence of strong ethical leaders may be integral to successful cooperation. This study by Sadowski, et al. played out a common good social dilemma among college students (2015). Experimenters, monitoring online chats between groups, found that there was a minority of reliably ethical students who shaped the decisions of the whole and ultimately the outcome of the game. The authors claim this as the study’s important finding, but some aspects of the results suggest that the participants were not taking the experiment at all seriously, which taints the integrity of the study’s findings. However, the idea that an influential few are integral to the behavior of the whole is consistent with findings in social psychology on early movers.

4.0.9 INCOME AND RISK DISPARITY

Income and risk inequalities within a game decrease cooperation. Yet participant income heterogeneity mirrors the demographics of the real world, suggesting that one of the barriers to cooperation and climate action is income and risk inequality. It appears that this occurs because low-resource/high-risk participants are less able to contribute, and high-income/low-risk participants are less willing. In a study by Calzolari, et. al, the real-world climate cost inequality was represented by a poor and a rich condition. On average, the rich group emitted less than the poor group did, although by small margins (averaging 6-15%). This detail is largely irrelevant in the more complex context of the real world, in which it is well known that socioeconomic status is correlated with higher emissions. In the tightly controlled game environment, rich participants were more able to contribute, and thus contributed more. However, across 3 studies of this type, a significant correlation between groups who successfully cooperated and groups where the rich participants emitted less was only found in one (Calzori, et. al, 2018).

Hurlstone, et. al, report a variation of this finding, in which rich participants only compensated for poor participants when all participants faced the same level of risk. In the four-condition study, when all groups had the same wealth and faced the same level of risk, 88% reached the game’s necessary threshold. When there was unequal wealth but equal risk, 63% reached the threshold, when there was wealth inequality but the rich were at higher risk, 75% reached it, but when there was wealth inequality and the poor were at risk, only 13% of groups reached the threshold, as the rich participants no longer had an incentive to help (hurlstone, et. al, 2016). The findings demonstrate that participants only consistently cooperate when they face risk themselves and have the capacity to avoid that risk. In the real world, lower-income groups are at higher risk, even though they are generally much lower emitters. The challenge that inequality poses to cooperation and successful mitigation must be solved.

4.0.10 SANCTIONS

The literature shows that sanctions reliably increase cooperation, especially deterring free riding when others are cooperating (Hurlstone, et. al 2017). Sanctions are punishments that players can impose when other parties are not cooperating. A “tit for tat” sanction is when one party’s deflection from cooperation is met with the same behavior as punishment. In common good games, this behavior is not helpful overall because it ends up exacerbating harm for all. It’s best when the cooperative party/ies can sanction the non-cooperative party in another way. Instead, disadvantages should be placed on the non-cooperative player. In global climate politics, this looks like trade sanctions or tariffs. According to Hurlstone, et. al, in order for these imposed punishments to work, they must be “severe and credible” (hurlstone, et. al, 2016). If the threat of sanctions is strong and well-believed, the sanctions themselves may never have to be put into action.

4.0.11 TIMEFRAME

Global warming is a massive crisis unlike humans have ever seen, which makes it difficult enough to conceptualize, but it also exists on a long and somewhat ambiguous time scale. Many individuals who take actions now do so to the benefit of generations beyond themselves; older generations will not live to see the most harmful effects of their emissions. Whether or not those alive now will see the harsher effects of global warming, it's still psychologically difficult to choose a lesser benefit in the present in order to receive a greater gain in the future. This is a concept called time discounting, a phenomenon often observed in economics, in which rewards in the future are psychologically treated as less valuable than in the present. This phenomenon is a concern because climate action in the present is meant to affect people in future generations.

The climate crisis’s extended timeframe poses a problem for mitigation. In a threshold cooperative game study on time discounting, players were less likely to cooperate the more distance there was between the present and their rewards. A standard cooperative game was posed in which there were three conditions: a short delay, long delay, and intergenerational delay condition. Participants were given, as usual, operating funds and an endowment. Their operating funds were used to play the game, and they were allowed to keep the leftovers that they didn’t use during the game. If they didn’t reach the game’s necessary threshold, there was a 90% chance that they would lose their endowment. They either received the endowment one day later, a week later, or it was invested into a reforestation project to benefit future generations. In the short delay condition, 70% of groups met the threshold, in the long-delay condition, only 36% met the threshold, and in the intergenerational condition, no one met the threshold (hurlstone, et. al, 2016). This suggests that highlighting the intergenerational responsibility of climate mitigation may not be an effective strategy. However, the study may have found a different effect if other factors were included or made salient, such as the individual human identity of members of future generations, or the potential that a sufferer of climate change could be a descendant. Additionally, the study had a pretty simple design that may not have truly communicated to the participants the significance of their contributions.

Fossil fuels will remain in the atmosphere for a very, very long time. This persistence is another heavy factor of the climate issue — cooperative behavior must happen in the context of persisting effects. However, the continuation of pollution may make cooperation less appealing or less effective over time. Because of this, Calzolari, Casari, and Ghidoni created a cooperative game in which participants were individually incentivized to produce emissions while it was simultaneously socially optimal to stay below a very low average level of emissions. The pollution costs of each round varied in different pollution persistence conditions. In the least persistent round, all of the effects of emissions were suffered within the one round, and then dissipated before the next round. In the intermediate condition, emissions persisted but decreased by half in each round. In the most persistent condition, all emissions persisted fully, cumulating until the end of the game. Participants’ rates of emissions reflected the pollution persistence of each condition. The no-consequence condition corresponded with the highest emission rates, likely because there wasn’t much of an incentive to cut emissions. In the intermediate persistence and full persistence conditions, the rates of emissions were lower than the no-persistence condition, but statistically insignificant from one another. This finding is somewhat encouraging, as the real world has a “persistent condition.” In the fully persistent condition, and only the fully persistent condition, 2% of groups were able to maintain the socially optimal level of emissions, but the vast majority of groups went well above the socially optimal level. The trend of average individual emissions in all groups went up over time (Calzolari, Casari, & Ghidoni, 2018). The persistence of pollution increased initial cooperation levels compared with the other condition, but as pollution accumulated/persisted throughout rounds, cooperation decreased. The authors suggest that this finding demonstrates the need to take extreme measures before cumulative effects increase to discouraging levels. It’s important to note that this cooperation game started with an initial pollution level of 0, which does not reflect the level of carbon pollution that already exists in the real world.

4.0.12 CONCLUSION

Throughout this chapter, it has been shown how cooperation is generally studied through cooperation games. However, it’s important to question whether these games really reflect how people will respond to the climate crisis. Do the results shown in these cooperative scenarios apply to the much, much broader context of the climate crisis? One of the ways that this has been addressed in the literature is through a paper by Claessens, et. al. This paper is a correlational study that looks at a so-called “cooperative phenotype,” but essentially tracks whether people’s tendency to cooperate in cooperative games is correlated with their self-reported frequency of environmental behaviors. There is in fact a correlation, which suggests that the people who are cooperating in cooperative games are actually representing people who are cooperating in real-world scenarios of the same character. This is a weak correlational study but is also one of the only things that speaks to this question (Claessens, et. al, 2022).

Another powerful question as to the integrity of cooperative games as a representation of the real-world collective action problem comes from a survey report by Tingley and Tomz. Although a lot of the discussion around the climate crisis and climate policy has regarded it as a collective action problem in which free riding is an issue, there is evidence to believe that the world’s citizens don’t see the issue that way at all (Tingley & Tomz, 2014).

In a survey report of the US and 12 European countries, Tingley and Tomz found that most people (an average of 79% across countries) want unconditional climate policy. In a survey of 16 developing countries, the average was even higher, 86%. This means that, regardless of how other countries behave, citizens believe that their country should continue to engage in mitigation action. This unconditionally cooperative attitude desires to mitigate climate change regardless of whether other countries are participating in climate action, with even some proportion thinking that their country should compensate when other countries are under-mitigating (Tingley & Tomz, 2014). In the context of a game theoretic framework, this kind of thinking and behavior is antithetical. However, not everyone is looking at the climate crisis like an economist, in terms of value gained and lost. It’s possible that citizens, in general, are approaching the crisis out of genuine concern from a value-driven standpoint, in which the point of fixing the climate crisis is not to get the most economic gain, but rather to save our species from immense pain, inequity, and even extinction.

4.0.13 BIBLIOGRAPHY

Alló, Maria, and Maria L. Loureiro. “The Role of Social Norms on Preferences towards Climate Change Policies: A Meta-Analysis.” Energy Policy, vol. 73, 2014, pp. 563–574, https://doi.org/10.1016/j.enpol.2014.04.042.

Brick, Kerri, et al. “Cooperation and Climate Change: Can Communication Facilitate the Provision of Public Goods in Heterogeneous Settings?” Environmental and Resource Economics, vol. 64, no. 3, 2015, pp. 421–443, https://doi.org/10.1007/s10640-015-9879-z.

Calzolari, Giacomo, and Marco Casari. “Carbon Is Forever: A Climate Change Experiment on Cooperation.” SSRN Electronic Journal, 2016, https://doi.org/10.2139/ssrn.2775276.

Constantino, Sara M., et al. “Scaling up Change: A Critical Review and Practical Guide to Harnessing Social Norms for Climate Action.” Psychological Science in the Public Interest, vol. 23, no. 2, 2022, pp. 50–97, https://doi.org/10.1177/15291006221105279.

Doherty, Kathryn L., and Thomas N. Webler. “Social Norms and Efficacy Beliefs Drive the Alarmed Segment’s Public-Sphere Climate Actions.” Nature Climate Change, vol. 6, no. 9, 2016, pp. 879–884, https://doi.org/10.1038/nclimate3025.

Farjam, Mike, et al. “Does Risk Communication Really Decrease Cooperation in Climate Change Mitigation?” Climatic Change, vol. 149, no. 2, 2018, pp. 147–158, https://doi.org/10.1007/s10584-018-2228-9.

Geiger, Nathaniel, and Janet K. Swim. “Climate of Silence: Pluralistic Ignorance as a Barrier to Climate Change Discussion.” Journal of Environmental Psychology, vol. 47, 2016, pp. 79–90, https://doi.org/10.1016/j.jenvp.2016.05.002.

Hurlstone, Mark J., et al. “Cooperation Studies of Catastrophe Avoidance: Implications for Climate Negotiations.” Climatic Change, vol. 140, no. 2, 2016, pp. 119–133, https://doi.org/10.1007/s10584-016-1838-3.

Irwin, Timothy. “Implications for Climate-Change Policy of Research on Cooperation in Social Dilemmas.” Policy Research Working Papers, 2009, https://doi.org/10.1596/1813-9450-5006.

Kelly, Daniel, et al. “Cooperative Phenotype Predicts Climate Change Belief and Pro-Environmental Behaviour.” Nature, vol. 12, 26 July 2022, https://doi.org/10.31234/osf.io/qu7v4.

Latkin, Carl, et al. “The Relationship between Social Norms, Avoidance, Future Orientation, and Willingness to Engage in Climate Change Advocacy Communications.” International Journal of Environmental Research and Public Health, vol. 18, no. 24, 2021, p. 13037, https://doi.org/10.3390/ijerph182413037.

Lübke, Christiane. “The Climate Change Dilemma: How Cooperation Beliefs Influence Energy Conservation Behavior.” Sustainability, vol. 13, no. 10, 2021, https://doi.org/10.3390/su13105575.

McEvoy, David M., and Todd L. Cherry. “The Prospects for Paris: Behavioral Insights into Unconditional Cooperation on Climate Change.” Palgrave Communications, vol. 2, no. 1, 2016, https://doi.org/10.1057/palcomms.2016.56.

Milinski, M., Sommerfeld, R. D., Krambeck, H.-J., Reed, F. A., & Marotzke, J. (2008). The collective-risk social dilemma and the prevention of simulated dangerous climate change. PNAS, 105(7), 2291-2294. https://www.pnas.org/doi/10.1073/pnas.0709546105

Sadowski, Jathan, et al. “Intergroup Cooperation in Common Pool Resource Dilemmas.” Science and Engineering Ethics, vol. 21, no. 5, 2014, pp. 1197–1215, https://doi.org/10.1007/s11948-014-9575-3.

Steg, Linda. “Psychology of Climate Change.” Annual Review of Psychology, vol. 74, no. 1, 2023, pp. 391–421, https://doi.org/10.1146/annurev-psych-032720-042905.

Sturm, Bodo, et al. “Conditional Cooperation in Case of a Global Public Good – Experimental Evidence from Climate Change Mitigation in Beijing.” China Economic Review, vol. 56, 2019, p. 101308, https://doi.org/10.1016/j.chieco.2019.101308.

Tingley, Dustin, and Michael Tomz. “Conditional Cooperation and Climate Change.” Comparative Political Studies, vol. 47, no. 3, 2013, pp. 344–368, https://doi.org/10.1177/0010414013509571.

Zelenski, John M., et al. “Cooperation Is in Our Nature: Nature Exposure May Promote Cooperative and Environmentally Sustainable Behavior.” Journal of Environmental Psychology, vol. 42, 2015, pp. 24–31, https://doi.org/10.1016/j.jenvp.2015.01.005.