Prudent cooperation

However, one difficulty remains. Mutual cooperation is the best solution for the group only if everyone cooperates. Only if all members identify with the group in the same way. But if I think as a member of the team whilst the others continue to think only of themselves, the result can be disastrous. I cooperate, the others do not, and I lose out completely. My attempt to achieve the common good produces a result that will be worse for me than the one I would have obtained if no one had cooperated.

Bacharach takes this objection very seriously. In *Beyond Individual Choice*, he develops a detailed version of team reasoning, which he calls ‘circumspect team reasoning’. The adjective is important because it introduces into the theory a consideration that we had hitherto overlooked. We cannot take it for granted that all individuals perceive a situation involving a choice in the same way. In other words, not everyone feels and reasons, for example, as a member of a team. For Bacharach, identification with the group is not a permanent characteristic of people. It is a psychological process that can be triggered by certain circumstances. In one situation I may regard myself as part of a team; in another, I may reason exclusively as an individual. And the same applies to others. This means that, when I have to decide what to do, I may be uncertain about the actions others will choose, due to the uncertainty inherent in the identification process. Will they be reasoning as individuals or as members of the group? Let us suppose, then, that each player has a certain probability of adopting collective reasoning. Those who reason as members of the team will try to carry out the collective plan, whilst those who do not will choose according to individual rationality. The question, then, becomes which plan is best for the group, taking into account the possibility that some of its members will not follow it. This is a different question from the one we have asked ourselves so far. We are no longer asking which combination of actions would yield the maximum benefit if everyone carried it out. We are asking which plan would yield the greatest expected overall benefit, whilst also taking into account the risk that someone might not do their bit. Bacharach shows that, in the prisoner’s dilemma, for example, which we have considered so far, the answer depends precisely on the probability (ω) of identifying with the group. If the probability that each player will follow the collective plan is sufficiently high, cooperation remains the best solution. But if that probability is low, then the situation changes. Cooperation yields the greatest benefit only if others cooperate as well. When this outcome becomes too unlikely, the risk of unilateral cooperation may outweigh the advantage of mutual cooperation. In such cases, choosing not to cooperate may even be consistent with reasoning geared towards the common good. Naturally, the threshold value of ω that makes cooperation rational depends on the values assigned to the various outcomes and on the model’s assumptions.

The conclusion that emerges from this type of analysis is profound and rich in practical implications. Simply wanting to cooperate is not enough for cooperation to automatically become the best choice. It also matters how likely we think others are to do the same.

Let’s return to the initial example of the shopkeepers and their street. Let’s suppose that some of them are genuinely convinced of the benefits of the regeneration project. They regard the street as a common good, recognise that the improvements would benefit everyone and would be willing to contribute, but the problem is that they do not know how many others share this conviction. Some might prefer not to take part, others might wait for their neighbours to foot the bill, whilst others might simply be sceptical about the initiative’s success. At this stage, a shopkeeper’s failure to participate tells us nothing about the value they attach to the project. They might be very keen for it to go ahead whilst, at the same time, considering the risk of being one of the few to fund it to be too high. Others, however, might simply not believe in the initiative at all.

Taking this possibility into account radically changes the way in which we traditionally interpret many problems of cooperation. We are accustomed to explaining a lack of participation in terms of selfishness, extreme individualism, indifference or a lack of prosocial motivation. Bacharach’s theory suggests that, at least in some cases, the problem may lie elsewhere. People may share a collective goal without believing it sufficiently likely that others will contribute to achieving it. This distinction is significant because it has very different implications for public policy. If the problem were merely selfishness, we would need to try to change individual preferences or incentives. But if the problem also involves uncertainty about the behaviour of others, then we could intervene to create conditions that increase the likelihood of widespread cooperation.

Let us consider a traders’ association that decides to launch the project only after having gathered a sufficient number of binding commitments. Or a system in which contributions are paid into a common fund and used only once the necessary sum has been reached; otherwise, they are refunded. In both cases, there is no need to alter the value that each shopkeeper attaches to the road improvement. It is sufficient to reduce the risk that their contribution will remain isolated – and be lost. This is not an individual issue, but an institutional one. We must not change the motivations of individuals, but the rules governing how they work together. Of course, these mechanisms do not eliminate all problems of cooperation. Differences in the assessment of benefits, distributional conflicts or opportunistic behaviour may still remain. These mechanisms address a specific aspect, making the participation of others more credible and, in this way, altering the conditions under which each person makes their decision.

Bacharach’s perspective thus allows us to view trust not only as a moral disposition, but also as a structural and epistemic condition that influences the possibilities for cooperation. If I consider it likely that others are willing to contribute to a common project, I may find it rational to do so myself. If, on the other hand, I believe that most others will not participate, the same choice may become too risky. The fundamental innovation is that this assessment may depend not so much on our beliefs about the nature of others’ motivations, but on our assessment of the probability that they will or will not adopt the ‘we’ perspective. Institutions, then, could operate not only by seeking to modify incentives through rewards, sanctions and controls, but also by attempting to reduce uncertainty about others’ behaviour, making commitments verifiable and helping to create the conditions in which collective reasoning becomes a practical reality. Consider the funding of public services, the management of environmental resources or the organisation of work. In all these areas, the success of an initiative often depends on the participation of a large number of individuals. Individual willingness to cooperate may be high and yet insufficient when the conditions that make others’ participation credible are lacking.

This observation also has implications for the way we assess individual responsibility. A person may be willing to contribute to the common good, but may not have sufficient resources to bear the risk of failure. Furthermore, the cost of this uncertainty is not the same for everyone. A company with ample financial resources can afford to participate in a collective project even when the outcome is uncertain. A small business might not be able to afford to do so. A worker in a secure position can bear the costs of a collective action that a worker in a precarious position might consider unsustainable. This is not necessarily a matter of differences in people’s moral character. It may be differences in the material conditions that make it possible to take on the risk of cooperation. Here, the issue of justice arises once more. If the possibility of cooperating, in fact, also depends on the ability to bear the costs of a potential failure, then economic inequalities can translate into inequalities in the very possibility of participating in collective projects. It is not enough to ask everyone to play their part when the risk of doing so falls to a very different extent on individuals. A policy aimed at fostering cooperation must therefore consider not only the incentives at play, but also the way in which the risks associated with collective action are distributed. It may seek to reduce them, to share them, or to ensure that they do not place a disproportionate burden on those with fewer resources. Michael Bacharach did not set out to develop a theory of distributive justice and never directly addressed these issues. His contribution is more technical and more limited in scope, but no less relevant to the subject.

One fundamental question remains, however. The shift from ‘I’ to ‘we’ can help us overcome the limitations of individual self-interest, making possible forms of cooperation that we would otherwise be unable even to imagine. But what is rational for one group is not necessarily desirable for everyone else. An organisation, for example, can very effectively pursue objectives that are socially harmful, just as a community can strengthen its internal solidarity by excluding those who do not belong to it. The problem, then, is not merely to understand how a ‘we’ comes into being, but also where we draw its boundaries and what consequences this choice has for others. For the ability to act together does not necessarily coincide with the ability to act justly.