Cooperative duties of efficiency and efficacy

ABSTRACT

I argue that agents can have duties to cooperate with one another if this increases their combined efficiency and/or efficacy in addressing ongoing collective moral problems. I call these duties cooperative duties of efficiency and efficacy. I focus particularly on collective agents and how agents ought to reason and act in the face of global moral problems. After setting out my account, I argue that a subset of cooperative duties of efficiency and efficacy of collective agents are duties of justice in virtue of the roles these agents have taken up.

KEYWORDS:

• Collective duties
• cooperation
• duties of justice
• efficiency
• efficacy
• global justice
• group agency

1. Introduction

Suppose you have to rake the leaves in your backyard. Next-door, I am tasked with raking the leaves in my backyard at the same time. Shortly after we both start our tasks, the wind starts unexpectedly blowing heavily, making it difficult to rake and bag the leaves before they are blown away again. Neither of us could have anticipated this, but the green waste bins are picked up later today and postponing our tasks is not a viable option. If we were to try to fulfill our tasks separately, raking the leaves would take a long time and cost a substantial amount of energy for each of us. If we were to cooperate with one another, however, raking the leaves would cost considerably less time and energy for both of us. This would leave us with more time and energy to complete our other tasks and hopefully to enjoy the remainder of the day. Should we cooperate with each other in light of these potential benefits? Intuitively, it seems that we should. The benefits of spending less time and energy on the same task provide us both a weighty pro tanto reason to work together. Of course, this reason can be outweighed by other considerations. If cooperating involves substantial costs, for example, you and I really dislike each other and working together would ruin both our days, then perhaps we should not cooperate. But such costs notwithstanding, we both may have reason all-things-considered to cooperate with one another to enhance our efficiency.

Next, suppose you and I are also tasked with cleaning out the sheds in our backyards. Both our sheds are completely filled to the roof. Once we are halfway through, we both notice that there are a few items that are too heavy to be lifted alone. Assume that at the start we both had no way of knowing that these heavy items were there, as they were completely hidden at first. If we would each fulfill our tasks individually, our sheds will both be quite well organized, but still contain these somewhat inconvenient heavy items. If we were to cooperate, however, we could also remove these heavy items and both our sheds would be perfectly organized. The potential upshot of cleaning out our sheds gives us both a pro tanto reason to join forces. If there are no significant downsides and the upshot is worth the effort, then we both have reason all-things-considered to cooperate with one another to enhance our efficacy. We may even have reason to cooperate because it enhances both our efficiency and efficacy. For example, by cleaning out our sheds together, we can not only remove the heavy items but also save fuel by driving to the dump only once. Thus, it seems we can have weighty practical reasons to cooperate with one another when this enables us to complete our tasks in a more efficient or efficacious way.

Our world is riddled with problems seemingly insurmountable from an individual perspective: overfishing, global consumerism, global poverty, world hunger, global health issues, climate change, natural resource depletion, terrorism, international conflicts, and so on. Solutions to these global moral problems require various sorts of collective action. Given that moral reasons are a subset of normative practical reasons, could we say something similar when our tasks do not concern raking leaves in a backyard or cleaning out sheds but remedying collective moral problems? Many agents are already discharging duties and attempting to solve (parts of) these problems. My focus will be on collective agents such as institutions, international organizations, or corporations. In this paper, I explore whether and when agents can be morally required to cooperate with each other if this enhances the agents’ combined efficacy and/or efficiency in their ongoing efforts to remedy one or more moral problems. I call these cooperative duties of efficiency and efficacy. Let me explain why this is relevant.

One might think that if two agents can increase their combined efficacy and efficiency by acting together when remedying a problem, then this is what they ought to have done from the start. But this need not be true. Discharging a duty may take a reasonable amount of time, especially when we are concerned with large-scale ongoing problems. Both agents’ capabilities and circumstances may change over time. This does not only concern an agent’s own capabilities but likewise those of other potential cooperation partners. Therefore, what is objectively required of an agent, in light of all facts, may change over time. For example, in the raking leaves case, the sudden strong wind changes what each of us has the most reason to do, because our circumstances and abilities have changed to a certain extent. We are still able to do our tasks, but it takes much more energy and time due to the wind.

Moreover, in my view, an agent’s moral reasons are relative to the agent’s available evidence (Kiesewetter 2017; Scanlon 2008; Zimmerman 2014). When a morally conscientious agent is deliberating about what to do, she is concerned with understanding what kind of ‘oughts’ are generated by these evidence-relative moral reasons. Next, suppose an agent discharges her evidence-relative duty at t₁, it is possible that at t₂ the agent is still in the midst of discharging her duty but her epistemic standpoint has changed. For example, suppose I could not have known at the start that some of the items in the shed are too heavy to be carried alone or that you would clean out your shed at the same time and that we could save fuel by driving to the dump together. What I have reason to do all-things-considered changes when I am already carrying out my task. The same may hold for moral reasons.

Thus, not only agents’ capabilities and circumstances can change, but agents may also become more knowledgeable of the problem or solution, or other relevant available information may change. All these factors can change an agent’s moral requirements. The possibility of increasing the efficiency and efficacy via cooperation changes the token-duties of the relevant agents.¹ What they morally ought to do has changed, as more efficient and efficacious solutions to address the ongoing moral problem have become available over time. This need not be only in cases where the agents are now together able to do something they individually cannot (e.g. lifting the heavy items), but it may simply concern cases where cooperation enhances their efficiency (e.g. driving to the dump together). Of course, changing course may come with various costs and risks and these must be weighed against potential gains.

In light of these considerations, I am primarily interested in whether agents can be morally required to cooperate with one another when they are already addressing parts of global moral problems. I will focus particularly on how agents should reason about with whom to cooperate and which parts of the problem they should focus on. In Section 2, I set out various ways in which agents can be morally required to do their part of collective actions in order to address global moral problems. In Section 3, I argue that agents must target particular parts of a problem that they are able to address when they cannot address the entire problem (whether individually or jointly). In Section 4, I set out four parameters of reasoning with regard to whom to cooperate with and which parts of the problem to address. With these parameters in mind, I spell out the jointly sufficient conditions for a pro tanto cooperative duty of efficiency and efficacy. I argue that in some cases agents can be morally required to cooperate when this enhances their combined efficacy and/or efficiency in their ongoing efforts to tackle global moral problems. In Section 5, I highlight the relevance of cooperative duties of efficiency and efficacy for global justice. I argue that some cooperative duties of efficiency and efficacy are duties of justice in virtue of the role agents have taken up.

2. Duties and collective problems

To see the potential import of increases in efficacy and efficiency, consider that, besides individuals, another important set of global actors for addressing these problems are organized groups, such as states, institutions, international organizations, or corporations. I assume that these organized groups (potentially) qualify as moral agents in their own right: they can have moral duties and are fit to be held responsible for their behavior and attitudes (Collins 2019; Copp 2006; de Haan 2022; French 1984; Hess 2014; Hindriks 2018; Lawford-Smith 2015; List and Pettit 2011; Pauer-Studer 2014). I adopt Christian List and Philip Pettit’s (2011) account of group agency: In order for a group of individuals to constitute an agent, it must form, revise and act upon representational and goal-seeking states, while satisfying desiderata of rationality, and exhibit these properties robustly. The decision-making procedure is shaped by the group agent’s constitution, that is, ‘a set of rules, formal or informal, for determining how the inputs of individuals are to be put together to generate group judgments as outputs’ (List and Pettit 2006, 89). To have a moral duty, an agent must be morally competent: the capacity to understand and process moral reasons and subsequently act accordingly. Due to their organizational make-up and decision-making structures, most group agents are capable of understanding and processing moral reasons in a functional sense akin to persons and can act accordingly (de Haan 2022). It is worth stressing that not all group agents necessarily qualify as a collective moral agent, some may lack moral competence (Hindriks 2018). For example, some failed- or quasi-states may fail to qualify as a moral agent (Erskine 2001). Still, many organized groups do qualify as moral agents and are often already discharging particular duties related to global moral problems.

In order to explain how (collective) agents can be morally required to cooperate if this increases their combined efficiency and efficacy in addressing ongoing moral problems, it is helpful to consider first what sort of problems agents might be confronted with and what sort of remedies these problems may demand. Let me therefore briefly explain the background framework to see what sort of duties agents can be discharged in relation to collective problems. Collective problems concern cases where some morally weighty outcome must be generated or prevented, but no agent has direct control over the outcome in the form of actions that are (necessary or) sufficient to prevent or generate that outcome. We can distinguish between two types of collective problems. Each type determines what is required of the agents, that is, the content of the correlative duties.

Contribution problems concern cases we typically find in rational choice theory where a weak sense of collective action is employed. Here concurrent individual actions considered together count as collective. What is morally required of the (unorganized) group is performing a particular collective pattern of actions that is (made) salient. Each agent has a duty to do or contribute her part. I do not make any claims about the nature of this requirement that befalls the group, it may be a conjunction of individual duties or a collective duty of some kind. What is important is that these contributions can be made in isolation from one another. A subset of contribution problems are collective harm problems, but contribution problems may equally concern sets of actions where the morally desirable upshot involves some benefit. To give a few real examples involving individuals: donating money to poverty relief (Cullity 2004; Singer 1972); refraining from buying factory-farmed chickens (Kagan 2011; Nefsky 2012); or not (over)using polluting modes of transport (Lawford-Smith 2016; cf. Sinnot-Armstrong 2005). But there are also contribution problems involving collective agents, for example, nuclear disarmament among states, moving to renewable energy among corporations, large-scale natural resource depletion, or reducing CO2 emissions among states. To discharge their duty to contribute, each actor can act in isolation from one another (of course, the actors may have relevant beliefs about mutual dependency, and so on).

However, some collective problems require agents not just to coordinate their actions but to cooperate and act together in a stronger sense in order to reach the morally desirable outcome. I call these cooperation problems. It is helpful to distinguish between shared actions and group actions. A group action is an action performed by a group agent in List and Pettit’s sense. But arguably a group need not be a group agent in order to act jointly (Schweikard and Pettit 2006). For example, suppose you and I are walking together down Fifth Avenue. We are not a group agent on List and Pettit’s account (cf. Gilbert 2014; Schmid 2018) and yet this is significantly different from walking alongside a stranger down Fifth Avenue at the same pace without bumping into each other. Both involve a sequence of individual coordinated acts, but the former case is importantly different. It is a shared action because of the participants’ shared (joint, or collective) intention to act together this way (Bratman 2014; Tuomela 2013). Without spelling out the details, I will assume that agents can act together this way without thereby constituting a group agent.

Cooperation problems can require either a shared action or a group action (by a group agent that is to be formed) for its remedy. Let’s start with shared actions. In some cases, agents are immediately able to perform a shared action that addresses the cooperation problem. For example, you and I are jointly able to subdue together an attacker that is beating up a victim, and it is clear what needs to be done. In other cases, agents must first engage in plan-like deliberation in order to perform the shared action that remedies the cooperation problem together. For example, a person is stuck under a collapsed building, we must lift various heavy blocks together in order to save the person, but it is not clear from the outset what we should do first. In both cases, the agents are morally required to act together, that is, to perform a shared action (cf. Held 1970; Pinkert 2014). How we should exactly explicate this claim is an important question. The debate concerns whether the relevant duties are individual (Collins 2019; Goodin 2017) or collective (Björnsson 2014; Cullity 2004; Pinkert 2014; Schwenkenbecher 2021; Wringe 2010). While this is a very interesting question, I believe the account I set out here is in principle compatible with all these views, so I remain as neutral as possible here.

In other cases, the agents must necessarily form a collective agent in order to address the cooperation problem. For example, a group of farmers must protect their town against a large wildfire using a strategy involving their agricultural equipment (irrigation planes, sprinkler systems, etc.). Suppose that the changing weather continuously requires changes in their plan such that in order to adequately and reliably protect the town from the wildfire, the farmers must organize themselves robustly and form a collective agent. In such cases, following Stephanie Collins (2013), the agents have collectivization duties, which are duties to form a collective moral agent (CMA) that can reliably and adequately address the morally pressing matter. As collectivizing itself is a shared action (Pinkert 2014; Schmid 2018), I leave open here whether collectivization duties are to be explicated in terms of individual or collective duties. The to-be-formed-CMA incurs a group-level duty to perform the remedying group action. To determine and perform the remedying group action, the group must adopt a decision-making procedure and distribute member-roles. Each member subsequently incurs an individual duty to perform her assigned member-role.

Note that for both contribution and cooperation problems, the relevant duty-bearers can be individuals or group agents. When concerned with group agents, the analysis simply moves up a level. In principle, given that group agents are agents with their own intentions, nothing stands in the way for a number of group agents to act together. The same holds for forming a (larger) group agent. When collective agents have collectivization duties, this need not necessarily imply that the collective agents must be dissolved or merged into one larger collective agent. This larger group agent may have a constitution of its own, meaning it has its own rules and procedures for facilitating group decisions and group action. The member-group agents ought to do their parts. For example, consider the United Nations, states are members but continue to be group agents with their own representational and goal-seeking states. Their interests may or may not align with the larger group agent, just as my goals, beliefs, or interests may or may not align with the university I work for.

To summarize, when facing contribution problems, agents must coordinate their actions and have duties to perform individual isolated actions. When facing cooperation problems, agents must cooperate with each other and have duties either to act together or to form a group agent that can address the problem. In what follows, I only focus on cooperation, although what I say here could potentially be relevant for coordination as well.²

3. Subdividing the problem

From the standpoint of most individual and collective agents, global moral problems are daunting. Most global moral problems are of such scale and complexity that finding sufficient willing cooperation partners to address the problem in full is either practically speaking impossible or overly costly in terms of time and resources. In such cases, it is best to focus on a smaller part of the problem. To spell out when agents exactly have cooperative duties of efficiency and efficacy, I must first set out how agents can target particular parts of a collective problem. This will be important for identifying potential cooperation partners in order to enhance efficiency and efficacy, as this will partly depend on what part of the problem they focus on.

If it is not feasible for individual or collective agents to address the collective problem in full via coordination or cooperation, then agents must subdivide the problem into sizeable parts, specifying and narrowing down the remedying collective pattern, shared action, or group action. Similarly, it makes sense to view certain global moral problems as a cluster of smaller yet interconnected problems. Subdividing the problem allows for creating sizeable parts that agents could address via various forms of coordination or cooperation. One can subdivide either compartmentally or geographically. One can specify the problem to a sub-problem, essentially compartmentalizing the problem. Or agents can subdivide the problem geographically, specifying the problem to a particular continent, region, country, or area. I call such a smaller subdivided part of the larger problem a subdivision.

For example, consider global poverty. Given the complexity and sheer magnitude of the problem, this is best seen as a set of collective problems. Global poverty does not simply concern people having insufficient monetary resources for subsistence but a series of systemic interconnected underlying problems. Consider the following non-exhaustive list of causes of persistent poverty: combatting the lack of economic growth alone is not sufficient to reduce poverty in households, as income inequality must be reduced for the poor to benefit. Stable environments are essential for the poor to climb out of poverty. Unrest together with corruption, dysfunctional bureaucracies and/or oppression and discrimination make the provision of basic services (to all) very difficult. The lack of (safe access to) health care, education and/or social security all ‘contribute’ to the pervasiveness of poverty. Hunger and poverty traps show how malnourishment and a lack of assets can cause households to be persistently stuck in hunger and poverty. International factors, such as unjust trade regimes, outside (military) intervention and interference can perpetuate or exacerbate poverty (Ahmed et al. 2007, 60).

Assuming it is not feasible (or morally desirable) to create a single agent to address global poverty in its entirety (e.g. a world state), agents should subdivide the problem of global poverty compartmentally and, for example, focus on the lack of basic services. If still too sizeable, agents can also geographically subdivide the problem, specifying the targeted problem to a particular continent, region, country or area. For example, agents may jointly target the lack of health care in a country. If the problem is still too sizeable, agents could target an even smaller part of that sub-problem, say focusing on combatting malaria in a country.

The general idea behind subdividing the problem is that when an agent is confronted with a problem where the objects of moral rights are at stake, but the agent cannot solve the problem directly in its entirety, then it is reasonable to expect the agent to start with a smaller issue if this potentially allows one (or others) to solve the entire problem, especially if solving this smaller issue is worthwhile in itself. Note that targeting a particular subdivision itself will safeguard the objects of a smaller set of rights. The agent’s moral reason to address this part of the problem, then, is that doing so safeguards the objects of a set of moral rights that are currently threatened.³ Let me stress that the idea is not that agents should simply pick any subdivision at will. The agents ought to select the subdivision they have most moral reason to address. In the next section, I will offer four parameters for agents’ reasoning about how to best subdivide the global moral problem when the agents are already discharging related duties, but these parameters are applicable to selecting a subdivision at the start as well.

Which subdivision the agents ought to target will also determine what sort of action is morally required. A subdivision can be a contribution or cooperation problem and may require a collective pattern of actions, acting together or collectivization for its remedy. In the absence of relevant organized groups, it is possible that for a group of individuals the best way to adequately and reliably address the largest feasible subdivision is via collectivization. For example, individuals may create an INGO to combat malaria in a particular country. Of course, this may easily be overdemanding. It is quite possible that there are collective agents (INGO’s) addressing important subdivisions that are the result of supererogation rather than the discharge of collectivization duties.

However, group agents are importantly different from individuals as duty-bearers. Once these collective agents are established, demandingness becomes less important. While we cannot reasonably demand any individual to give up their entire life, this does not hold for these collective agents. We can demand more from these group agents not only comparative to their abilities but also given the sort of agent they are. As Collins (2020) points out, collective duties cannot be overdemanding. Instead, collective demandingness complaints should be reinterpreted as complaints about the inability or third-party costs. Moreover, and perhaps more importantly, the nature of the moral requirement of these collective agents can change. Unless an individual is connected to a subdivision in a particular way, an individual’s duty might be ‘merely’ a duty of beneficence.⁴ But, as I will argue in Section 5, the collective duties of these collective agents are duties of justice in virtue of their assumed roles to protect and safeguard the objects of a set of moral rights.

It is worthwhile to see that the presence of such organized groups allows for individuals to target different subdivisions, and so the presence of group agents affects what individuals have most moral reason to do. For example, INGO’s may enable individuals to make isolated contributions (or to simply join the group agent) to a particular subdivision that otherwise they could not have made (e.g. donating money to poverty relief). But in some cases, the best way to ensure a particular subdivision is addressed is via political action rather than making isolated contributions, and individuals have duties to act together to place pressure on particular group agents to (help) remedy this problem. It is possible that for some global moral problems, climate change for example, agents are required to do both (e.g. flying less and demonstrating). How an agent is to reason about all these obligations is beyond the scope of this paper, but within constraints of demandingness, individuals are required to act in various ways in relation to various subdivisions of (various) global moral problems.

Note that few (if any) global moral problems can be fully addressed without the involvement of various group agents. Group agents can potentially target large subdivisions of global moral problems on their own and are often already discharging relevant duties related to particular subdivisions (or the entire problem). For example, most states have foreign aid policy including bilateral and multilateral aid. The UN’s sustainable development goals target climate change, global poverty, and its underlying causes. The World Bank aims to end poverty by creating sustainable economic growth, investing in people, and building resilience to shocks and threats. INGOs all target specific sub-problems of global moral problems.

However, given that it is practically unfeasible to know and account for all relevant factors from the get-go, as there are simply too many underlying causes and relevant factors to consider, it is likely that an agent’s knowledge of the problem and available evidence improves over time. Moreover, the best strategy is also dependent on other agents’ abilities and actions, and other circumstances related to the general problem. Hence, even if all agents at some point subdivide the problem in the best way possible and decide to discharge the exact right duty with the exact right strategy at that time, what the best strategy is may change over time. With this in mind, I turn now to the question whether group agents can be morally required to cooperate with other group agents if this increases their combined efficiency and/or efficacy.

4. Efficiency and efficacy

Increases in efficiency matter because this enables agents to reallocate time and resources and reorientate their efforts to new important projects directed at solving other (related) moral problems. Increases in efficacy matter because this means that more people’s objects of rights are secured. To make it intuitive that group agents can have moral duties to cooperate if this enables them to increase their efficiency and efficacy, consider the following case.

Fire Brigades: A large wildfire moving south threatens the lives of many inhabitants of a region. Fire brigade Alpha is combatting the fire Southeast to protect city A. Fire brigade Bravo is protecting city B in their district Southwest. However, in the middle is a smaller town C that is not protected. Without further assistance inhabitants of town C are likely to die. Individually, neither fire brigade is able to protect both their city and C. If Alpha and Bravo were to bundle forces, allowing them to use their manpower and resources more efficiently, each withdrawing a few surplus firemen and trucks, they could adequately protect towns A, B and C from the wildfire.

Prima facie, fire brigades Alpha and Bravo ought to act together if this enables them to save all the inhabitants. This gives us good reason to think collective agents can have cooperative duties if this enables them to better solve one or more moral problems.⁵ The fact that the collective problem they are facing, a situation in which numerous objects of rights are threatened, can only be addressed via cooperation is a weighty pro tanto moral reason for each fire brigade to cooperate. This need not only concern duties to act together. Suppose that the wildfires become a continuous threat and the fire brigades must organize themselves robustly to be able to protect all three towns adequately and reliably from possible new wildfires. For example, suppose that a timely response to new wildfires requires a specific distribution of equipment and manpower, but without robust organization, the right distribution cannot be guaranteed, and the group is not able to reliably protect all three cities. If so, the fire brigades have duties to collectivize and the newly formed group agent has a collective duty to reliably protect all three towns. Thus, agents can in principle have cooperative duties to either act together or collectivize in order to enhance their combined efficiency and/or efficacy, although in practice this is more likely to concern acting together than collectivizing.

As I mentioned, subdividing the problem may be done geographically and compartmentally. Cooperative duties of efficiency and efficacy may apply to agents targeting different geographical subdivisions. Quite recently Singapore-based ASEAN Macroeconomic Research Office (AMRO) and the International Monetary Fund (IMF) pledged to enhance cooperation to respond more effectively, thereby strengthening their partnership in promoting regional and global financial stability.⁶ Suppose AMRO is attempting to improve the economy of country A, both in terms of GDP and income inequality. The IMF is likewise trying to increase the economic development and income inequality in A’s geographical neighbor country B. Both organizations have combatting severe poverty as their intended aim. Suppose that by cooperating and coordinating their efforts, they can foster mutual beneficial trade between A and B, but also include country C, another neighboring country. At the same time, they could share know-how and expertise and achieve economies of scale. They would both be realizing their initial aim but increase their efficiency as well. Moreover, they would increase their efficacy, being able to assist three countries instead of two. This is a simplified example, of course, but the point should be clear.

Cooperative duties of efficiency and efficacy equally apply to agents working on different components of a moral problem. Suppose an INGO aims to install reliable water supplies in central parts of Africa to combat global poverty. Suppose they could increase their efficiency and efficacy by cooperating with an INGO that aims to empower farm holders in developing countries via agricultural projects that increase food production. It is not hard to recognize the mutual beneficial relation between a reliable water supply and increases in agricultural productivity. The reduction of (transaction) costs, and sharing expertise, know-how, and resources would result in increases in efficiency. Both INGOs would also increase their efficacy because the reliable water supplies are now used for multiple purposes and the agricultural projects are supported by a reliable water supply.

Let me stress that fostering increases in efficiency and efficacy does not merely concern ‘donor-coordination’. What I mean by this is not just that aid-recipients are accountable (cf. Drydyk 2019), but aid-recipients can be among the moral agents who have cooperative duties of efficiency and efficacy. This need not just concern states or regimes, but may concern local authorities, other group agents, and even individuals in some cases. The scale of the subdivision will likely determine in part who will have the correlative cooperative duties, because the agents must be able to contribute to addressing the problem. For example, an INGO may have cooperative duties together with locals, whereas the IMF is more likely to share cooperative duties with states or regimes simply in virtue of the type of subdivision they ought to address.

Of course, it could be overdemanding for some aid-recipients to have such duties. In such cases, there are nonetheless good reasons to include aid-recipients in the cooperation. One good reason is that it may enhance the efficiency and efficacy of aid. If a group of individuals is trapped in a burning house, calling out for the fire brigade to find their position faster, should the fire brigade ignore their calls and continue their rescue option as they see fit? No, we expect them to adjust accordingly and make use of the inside knowledge. Similarly, alignment and ownership may be key in some cases for increasing the efficiency and efficacy of assistance, just as coordination with local authorities might be essential for humanitarian interventions in case of natural disasters. If including aid-recipients allows for increases in efficiency and efficacy, they ought to be included.⁷ But even if including aid-recipients does not necessarily increase efficiency or efficacy, when feasible, they ought to be included in order to respect their agency and to enhance their wellbeing and equity (cf. Drydyk 2019).⁸

Next, I turn to how agents ought to select which subdivision to target and with whom to cooperate. The following parameters should play a role in an agent’s reasoning:

Efficiency and Efficacy: The relevant set of subdivisions to consider contains those subdivisions where the agent can best increase the combined efficiency and/or efficacy with other potential cooperation partners. One particular subdivision may have various sets of possible cooperation partners, each combination has its own respective gains in combined efficiency and efficacy. The number of cooperation partners is important. Too large potentially allows for inefficiencies, but too small could mean possibilities are left untouched.

Costs: Shared action or forming a group agent comes with various sorts of costs, which need to be offset against potential gains in efficiency and efficacy. While some costs directly lower efficiency and efficacy (e.g. the costs of shared action or collectivization itself), other costs are indirect. For example, agents may already have spent considerable resources to facilitate their current remedying action, these investments should be preserved as much as possible. It may also take valuable time and resources to convince other cooperating partners to join in, whereas targeting a different subdivision with the same increase in efficiency/efficacy need not involve such costs. There are also moral and political costs that must be taken into consideration. For example, it may be morally problematic if two particular agents collectivize. And perhaps agents should avoid cooperating with states or regimes that are deeply unjust, immoral or corrupt. Or a particular subdivision may require agents to violate other (special) obligations, whereas another subdivision does not.

Connection: Agents must take morally relevant connections to particular subdivisions or other potential cooperation partners into consideration. Some actors may have duties to rectify previous harms or injustices, and this gives them stronger reason to target particular subdivisions.⁹ For example, the agent (partially) caused the harm or injustice or perhaps in some cases benefited from an injustice. This may require agents to target another subdivision with less gains in efficiency and/or efficacy or in some cases even trump all considerations of efficiency and efficacy. Some actors have special obligations towards particular agents, this can affect what subdivision to target or with whom to cooperate. These special obligations can be related to moral or political costs. For example, a state has specific obligations towards its citizens and violating these obligations involves moral and political costs that may trump potential increases in efficiency and/or efficacy. If such considerations trump a cooperative duty concerning this particular subdivision, then the agent should select the next best possible subdivision where such costs are less or absent.

Success: The envisioned group (agent) must be capable of performing the remedying shared or group action related to the selected subdivision. The probability that the group (agent) successfully performs the remedying shared or group action, φ, must be over some threshold (Collins 2013, 239). This threshold is determined by the context. In this context the (moral) importance of φ is included, the more important φ, the lower the threshold. For example, think of deterring a giant asteroid from hitting Earth (Collins 2013, 240), the importance of the problem significantly lowers the threshold. But equally important is whether a failed attempt to φ possibly has bad consequences other than not φ-ing itself. If so, the threshold is raised. Moreover, the fact that the agents are already discharging duties raises the threshold. Agents should not be better off continuing their current efforts, because the likelihood of φ-ing is low. Finally, the costs of cooperating influence the threshold as well, the higher the costs the more probable φ must be.¹⁰

A fine balance needs to be struck between Efficiency and Efficacy, Costs, Connection and Success when selecting the subdivision and cooperation partners. Agents must not overreach, collectively taking on a problem bigger than they can handle. But they neither must waste potential increases in efficiency and efficacy, had they targeted a different subdivision, even if φ’s likelihood would be lower. Holding Costs and Connection fixed, agents ought to target the subdivision where efficiency/efficacy is as high as possible, but where the collective would still meet the threshold. How to weigh Costs or Connection compared to increases in efficacy or efficiency will depend on the specific costs or connections involved, but this must already be taken into consideration when selecting the subdivision and not when considering whether one has a cooperative duty all-things-considered, because an agent might overlook particular subdivisions where it would have a cooperative duty all-things-considered.

I can now formulate the following jointly sufficient conditions for a pro tanto cooperative duty of efficiency and efficacy:

If

1. an agent can increase their combined efficiency and/or efficacy in safeguarding the objects of a set of moral rights by cooperating with one or more agents, and

2. jointly or collectively targeting a subdivision of the general problem would enable this increase, and

3. there is at least one potential group (agent) where the probability of a successful remedying action is over the relevant threshold for which Cost and Connection considerations do not outweigh Efficiency and Efficacy and Success considerations,

then the agent has at t₁ a pro tanto cooperative duty to:

1. at t₁, identify the best combination of (a) targeting a subdivision of the problem and (b) potential cooperation partners given Efficiency / Efficacy, Costs, Connection and Success, and

2. at t₂, take steps towards and prepare for cooperating with the identified potential cooperation partners, and

3. at t₃, depending on what sort of action is required to remedy the selected subdivision, perform their part of (a) the shared action or (b) forming a group agent that can remedy the selected subdivision.

Note that this is a pro tanto duty. This means the duty can be overridden by other stronger duties in cases of conflict. As formulated here, the pro tanto cooperative duty of efficiency and efficacy is individual, but the account can easily be amended such that this individual duty is part of a collective duty to act together or jointly form a group agent.

To select a particular subdivision and potential cooperation partners is a demanding epistemic task, as agents should consider various probabilities, costs, connections for various possible subdivisions and potential cooperation partners. Of course, each agent ought to do so to the best of their ability. Admittedly, this may potentially be quite demanding for individuals. But given the roles of particular group agents, it is not unreasonable to expect group agents, who are already addressing (parts of) a particular subdivision, to routinely check for potential better solutions related to that subdivision. Notably, there may be differences between the available evidence of two agents, such that the more-informed-agent has already a pro tanto cooperative duty at t₁ whereas the less-informed-agent does not. At t₂, the more-informed-agent ought to inform and mobilize the other potential cooperation partners, and the less-informed-agent’s available evidence will be updated. At t₃, then, the agents ought to perform their part of the remedying shared action or formation of the collective agent that incurs a duty to address the selected subdivision.

5. Cooperative duties of justice

So far, I have argued that agents can have pro tanto cooperative duties of efficiency and efficacy related to global moral problems. Given the focus on agents’ duties to perform actions, this account falls under what some call an interactional approach to global justice. The interactional approach is usually contrasted with an institutional approach that focuses on distributive justice and institutional structures (Beitz 1999; Pogge 2002; Tan 2004). The main criticism is that interactional approaches overlook important questions, for example, questions concerning global wealth and resource distribution, and the structure of international political and economic institutions (Boran 2008). I agree that to frame issues of global justice solely in terms of humanitarian and development aid is a mistake, and that duties of assistance do not fully address all relevant matters of global justice (cf. Rawls 1999). But, without doing justice to this intricate debate, let me just note that changes in the background structures must be implemented by agents, and that agents, if needed, can create new overseeing agencies to enforce particular (redistributive) policies, patterns or background structures. There is no reason to think the interactional approach is limited to the interactions between states, as there are many other global institutional actors. We can see the interactional approach as part of non-ideal theory that should be informed by and aspire to approximate the (partial) ideality of the best institutional approach.¹¹ In my view, a non-ideal interactional approach must take into account that (1) agents often do not possess all relevant information; (2) other agents may be non-compliant; (3) there is a scarcity of resources; and (4) agents have histories and connections that affect what they morally ought to do. The interactional approach spells out what actions are morally required of various moral agents under these non-ideal conditions.

Next, I’ll show that the interactional approach allows us to focus on an important part of global justice that is overlooked on a purely institutional approach. I’ll show that a subset of cooperative duties of efficiency and efficacy are duties of justice in virtue of the role these group agents have taken up. I rely upon a rather stringent view of duties of justice, but I think the argumentation below is compatible with more lenient views as well (Miller 2007; Singer 1972; Valentini 2013).

What makes a moral requirement a duty of justice rather than a duty of beneficence? Duties of justice specify what is owed to those with whom we stand in relations of justice and these agents have a justified claim-right to the (non-)performance of such actions.¹² Because of this, there is good reason to think that an agent is under a duty of justice only if the duty is perfect (Kant 1991; Mill 1998; O’Neill 1996), meaning its content and mode of performance are fully specified. All duties of justice are perfect duties, whereas all duties of beneficence are imperfect duties. Perfect duties concern the performance of actions that are owed to either all or specific agents. For example, one’s perfect duty not to make false promises is owed to all. But once I promise you something, I incur a promissory duty of justice owed specifically to you. The duty-bearer has a pro tanto reason to discharge them on every occasion on which they arise (Goodin 2017, 269). With imperfect duties, in contrast, you only have a pro tanto reason to discharge them on some occasions.¹³ This means that imperfect duties come with a certain latitude, that is, it is up to the discretion of the duty-bearer when and how to discharge them. Without perfection, it is not clear how others could have a justified claim-right to the (non-)performance of actions.

But not all perfect duties are duties of justice. Perfection is a necessary condition, but not a sufficient one. Duties of justice are only those perfect duties that are enforceable (Kant 1991; O’Neill 1996). This means that external coercion to comply with such a duty is morally possible (Kant 1991, 188). Compliance to these moral demands can be (in the normative sense) externally enforced by certain moral, legal or social sanctions (Tan 1997, 53). The state is usually the coercive apparatus invoked to protect rights and enforce these pre-institutional duties of justice, but other institutions or overseeing agencies can be created to do so as well. Other group agents are created and bestowed with particular roles, specific competencies, and duties necessary for the protection of these rights but still fall under the coercive apparatus of the state.

Group agents can also play a mediating role (Shue 1988), transforming imperfect duties into perfect ones (Beitz and Goodin 2009, 15). Duties of beneficence are imperfect in the sense that a particular act-type is required, but the act-token and particular occasion to discharge the duty is optional. The creation of some institutions can be seen as a manner of consolidating and perfecting pre-institutional duties of beneficence. As Robert Goodin explains, the idea is essentially this:

Start with a group of people who pre-institutionally, have imperfect duties toward one another. Then introduce an institution that (a) assumes all the imperfect duties of charity of all the duty-bearers within that group, and (b) acquires corresponding duties toward all of those who were supposed to benefit from those duties. (2017, 275)

To avoid the difficulty to match any particular duty-bearer to a would-be beneficiary, the institution has a duty to a number of beneficiaries as a whole. These group agents all have a positive duty of aid with regard to the purpose for which they were set up. In my view, these institutions can be seen as the result of the discharge of collectivization duties (see also Goodin 2017, 277).¹⁴

What is crucial is that many group agents with cooperative duties of efficiency and efficacy have all assumed a particular role. To explain this, consider an example discussed by Thomas Pogge. Suppose a police officer stands idly by when a homeless woman is beaten to death. One would expect an alarmed police officer to intervene because it is his role. Unlike ordinary bystanders, Pogge argues, this officer does not merely breach his positive duty to assist or protect like any bystander would, but additionally violates his duty to respect human rights (2011, 11). The officer violates a negative duty not to occupy a role and then fail to fulfill that role in a way that foreseeably and avoidably renders insecure access to objects of human rights. This overall duty is negative, like the duty not to make false promises, but once one has assumed the role of police officer, this may generate positive duties for further action (i.e. to protect citizens). Of course, there is an important difference between the police officer and collective agents who do not discharge their cooperative duties of efficiency/efficacy, as the latter are often already discharging other relevant duties. But Pogge is on to something here.

The crux of the story lies in the creation of a perfect duty of justice, just as making a promise does. Note that all these various group agents have assumed roles to address particular subdivisions, and to safeguard and protect the objects of a set of moral rights. Let’s call such roles: right-safeguarding-roles. The constitution of the group agent specifies the exact right-safeguarding-role it has assumed, namely, to consolidate a set of imperfect duties to an identified set of beneficiaries as a whole. Adopting a right-safeguarding-role creates a relation of dependency. The right-holders may reasonably expect the role-occupiers to protect the objects of their moral rights. Assuming and then foreseeably and avoidably failing to fulfill such an important right-safeguarding-role means that other agents will (likely) not (organize themselves to) assume this role. Had other agents done so and fulfilled it correctly, some people could (presumably) have led autonomous lives, because the objects of their moral rights would have been safeguarded. Neglecting what is required to adequately fulfill that right-safeguarding-role is a violation of a general duty not to assume a right-safeguarding-role and then fail to fulfill that role in a way that foreseeably and avoidably renders insecure access to objects of moral rights. This general duty is perfect because it is owed to all, but once one assumes such a role, much like a promise, this generates a specific perfect duty of justice that is owed to those one should protect.¹⁵

The point is that adequately fulfilling such a right-safeguarding-role may require further cooperation. Suppose the police officer witnesses someone being attacked by four assailants. Alone he cannot stop this particular crime, but he is certainly under obligation to call for back-up. Not doing so would be a gross neglect of one’s duty as a police officer to uphold the law and secure the rights of those protected by it. Similarly, the role of these group agents requires them to cooperate in order to increase their efficiency and efficacy. Of course, in Fire Brigades the collective agents are indeed already doing something. But more may be morally required of agents with such a role. The point is that if they (are able to) foresee that by cooperating they could secure safe access to either (i) more objects of rights of a particular set of beneficiaries; (ii) the same objects of rights of a larger set of beneficiaries; or (iii) both, then this is part of their right-safeguarding-role and neglecting that cooperative duty implies a violation of the general duty not to assume a role and then fail to fulfill that role in a way that foreseeably and avoidably renders insecure access to objects of rights.

One might object that this does not yet fully explain why the fire brigades’ cooperative duties of efficiency and efficacy are duties of justice, because it seems these group agents are created to protect town A and B respectively, and not town C.¹⁶ The same goes for the two INGO’s that target installing reliable water supplies and empowering farm holders in developing countries by employing agricultural projects that increase food production. If they must go beyond their initially targeted subdivision, how can this be a case of justice in virtue of their role?

It is important to see that from a normative perspective what is exactly required to fulfill the role specified in the constitution of these group agents is not set in stone. These group agents are not necessarily tied to a specific subdivision throughout time. The group agent is initially created to consolidate imperfect duties and remedy a particular subdivision of the collective problem, but this does not exhaust what the group agent ought to do. The group agent incurs a positive duty of aid to a determined set of beneficiaries, but the exact content of this duty may change. This is because the set of relevant beneficiaries can change over time, for example because more agents have joined and now support the group agent. The set of beneficiaries may also simply enlarge if the group agent becomes more efficient, because the duties of beneficence that generate the perfect duty are, after all, imperfect. Moreover, we know that the best (available) way to secure safe access to (specific) objects of any specific set of beneficiaries may change over time, as agents can become more knowledgeable and relevant circumstances may have changed. Hence, it is possible that after some time the best way to consolidate these imperfect duties is by cooperating with one or more (group) agents.

Interestingly, this points to a domain of global justice that is not the usual focus on either the interactional or institutional approach. Because these collective agents have assumed a role that is directly concerned with addressing these problems, this makes it a case of justice irrespective of the retrospective. We’re not concerned (necessarily) with rectifying previous harms or injustices. Besides backward-looking factors, the scope of distributive justice is equally irrelevant, because this is a different matter of justice altogether. Global justice also requires agents to adequately discharge the duties generated by their assumed right-safeguarding-roles. Due to the right-safeguarding-roles these group agents have assumed and the dependency of right-holders this entails, when the jointly sufficient conditions spelled out in Section 4 obtain, each incurs a cooperative duty of justice. These cooperative duties of efficiency and efficacy are duties of justice, then, because they derive from a general duty of justice not to assume a right-safeguarding-role, and then foreseeably and avoidably fail to adequately fulfill that role. All these duties are perfect and enforceable.

6. Conclusion

In this paper, I have argued that moral agents can have duties to cooperate with one another if this increases their combined efficiency and/or efficacy in addressing ongoing collective moral problems. When agents cannot address a collective problem in its entirety, agents must subdivide the problem either compartmentally or geographically. Over time, what exactly is morally required of moral agents may change. This is especially important for many group agents who are already addressing important subdivisions of global moral problems. From a moral point of view, it matters that these group agents have assumed various roles to protect the objects of a set of moral rights. When group agents with such roles can safeguard more objects of moral rights via cooperation, then these group agents’ cooperative duties of efficiency and efficacy are duties of justice.

Let me reflect briefly on what it means for it to be morally possible to enforce such cooperative duties of justice. In order to give practical meaning and effective implementation to duties of justice typically precisification through the institutional specification is needed to make such moral demands concrete and effective in real life (Goodin 2017, 272). For example, to enjoy, protect, and respect pre-institutional property rights, we need property laws to make all this precise enough to guide our day-to-day behavior (Goodin 2017, 273). Similarly, this general pre-institutional duty of justice not to assume right-safeguarding-roles and then foreseeably and avoidably fail to adequately fulfill that role and the special obligations generated by this general duty may require further precisification. But given that at the global level there is not (nor is it morally desirable to create) one single agent (that is, a world state) with coercive powers that can do so, the enforcement of such duties is something various group agents need to do together. Moreover, we may need to redesign our global institutional framework to some extent such that these duties can be effectively enforced. One plausible option is to encode this somehow in international law. What the right proportionality is for any coercive measures is a question I will have to leave open for now.

What I have shown is that some collective agent’s cooperative duties of efficiency and efficacy are duties of justice. This means global justice requires more than remedying past injustices or harms nor is it ‘merely’ about a just distribution. The upshot of this is that we may view the interactional approach as a manner of striving towards what is the best institutional approach while addressing relevant issues concerning global justice in non-ideal theory.

Acknowledgements

I am very thankful for helpful comments from Franz Altner, Stephanie Collins, Daan Evers, Christina Friedlaender, Frank Hindriks, Grace Paterson, Herlinde Pauer-Studer, Matt Rachar, Hans Bernhard Schmid, Michael Schmitz, Anne Schwenkenbecher, Leonie Smith, Bart Streumer, Leo Townsend, Elisabeth Widmer, Bill Wringe and several anonymous referees of Journal of Global Ethics. I am also thankful to audiences of the MANCEPT Workshop in Political Theory 2016: Expanding the Horizon: Collective Moral Agency and Global Justice at University of Manchester, and the Association for Social and Political Philosophy Annual Conference 2019: Justice in Times of Austerity, Rupture and Polarisation at Newcastle University.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program [grant agreement number 740922] (H2020 European Research Council).

Notes on contributors

Niels de Haan

Niels de Haan is a postdoctoral researcher and lecturer at the University of Vienna. His research is primarily focused on normative ethics and social ontology, in particular on collective moral agency, collective responsibility, collective duties, and normative reasons.

Notes

1 Some might think that what has changed is the best possible means for discharging the existing duty, but not the duty. This depends on how coarse- or fine-grained the act-token specified in the content of a duty is. In my view, this is fine-grained, meaning the token-duty has changed, because the agents ought to perform a different act-token. I am thankful to a referee for pressing me on this.

2 This would mean setting out contributory duties of efficiency and efficacy, which I will not do here.

3 I have in mind the objects of a set of moral rights that follow from one’s substantive ethical theory (e.g. contractualism or rule-utilitarianism).

4 Individuals may have duties to rectify prior injustices or to stop harming others, which are both duties of justice. If the best way to stop harming others or to rectify a prior injustice is via collectivization, then this yields collectivization duties (and a resulting collective duty) of justice.

5 Likewise, if the most efficient and efficacious way to address several moral problems requires splitting up an organization, then this may be what is morally required. I do not have the space to consider this in detail here.

6 For the press release of the IMF, see: https://www.imf.org/en/News/Articles/2017/10/11/pr17395-imf-and-the-amro-agree-to-enhance-cooperation.

7 Unless there are, of course, overriding reasons not to do so. For example, if regimes or local authorities are corrupt, it may be best to avoid including them in the cooperation (if feasible) in certain cases.

8 I am thankful to an anonymous referee of Journal of Global Ethics for pressing me on this.

9 Note that there is a possible link between cooperative duties of efficiency and efficacy and duties of redistribution according to the best institutional approach, for example Thomas Pogge’s global resources dividend proposal.

10 I do not mean to imply that these parameters are exhaustive.

11 I am not claiming this substantially differs from an institutional approach.

12 I wish to leave open here what this exactly entails, as this depends on one’s substantive ethical theory and how one exactly conceives of interpersonal relations of justice and what we owe to other agents.

13 As Goodin points out correctly, this does not necessarily make duties of justice more important than duties of charity in cases of conflict (2017, 270).

14 This need not only be true for individuals’ duties of beneficence. Group agents can likewise be created to consolidate rectificatory duties of justice of individuals, namely when this is the best way to remedy the injustice.

15 This conclusion follows on other views too. On Laura Valentini’s autonomy view, an agent is under a duty of justice if and only if she has the ability to refrain from undermining the necessary conditions for others to lead autonomous lives (Valentini 2013, 496). The agent has this ability if the agent is able to foresee the effects of their actions and has sufficient control. Duties of justice are correlative to a right not to be deprived of the social conditions to lead autonomous lives (spelled out in either egalitarian or sufficientarian terms). Given that these collective agents are able to foresee that cooperation is necessary for appropriately fulfilling their role and have sufficient control, they are undermining the necessary conditions for people to lead autonomous lives.

16 I am thankful to a referee for suggesting this objection.