The problem I am concerned with is very general: Why do we need a coercive institution in our society to control our behavior? This question is a little different from "Why do we need a government?" in two ways: First, because "coercive institution" is a broader term than "government"; probably not every coercive institution that controlled people's behavior would be called a government, though every government is a coercive institution (that is, an institution exercising coercion as one of its main functions). Second, controlling the behavior of the members of the society in which it exists is perhaps not the only important function of government (it may be, for example, that we need a government to fight people from other societies); I will not consider other possible things we might need a coercive institution to do. I will only consider whether we as a society need such an institution to control our own behavior.
Nor will I consider every possible, plausible reason we might need such a thing. I will consider only one reason, though it is an important one and, I think, the main reason why many people believe that we must have a coercive institution that controls us in certain ways. This reason is that such an institution, acting in such a way (roughly speaking), is necessary to maintain our own well-being, to any tolerable degree, and this because it is the only way to avoid a 'tragedy of the commons' situation, in the sense of that term that Garrett Hardin employs in his article of that name. How exactly this argument goes we shall see shortly. The reasoning depends on three claims: first, that there is a tragedy-of-the-commons problem, which is serious; second, that having a coercive institution to control our behavior solves the problem, at least reasonably well; and third, that no other solution would work tolerably well. This is what Garrett Hardin in fact argues in "The Tragedy of the Commons."
I will argue that social coercion (as I call it) is not justified, at least not for the preceding reason, both because it does not solve the tragedy-of-the-commons problem and because it is possible to solve the problem otherwise.
There are two main examples of 'the tragedy of the commons' that Hardin discusses in his famous article (along with several less important illustrations). The first is that from which the general situation-type gets its name: Suppose there are a number of cattle ranchers who each have access to a common grazing area. Assume that each rancher is free to add as many cattle to his herd and graze them on the land as he can fit in the limited area. Assume also that the commons has a certain optimum capacity for supporting cattle such that, if this capacity is exceeded, the commons will quickly be degraded due to overgrazing and become useless to everyone. No one wants this to happen; however, Hardin argues, each rancher is rationally compelled to put as many of his own animals on the land as possible. This is because he reaps all of the benefit from adding another animal to the herd, whereas the costs in terms of overgrazing are shared by everyone, so the individual rancher bears only a small fraction of the total costs created by his adding more cattle to the land. If, on the other hand, he chooses to limit the size of his herd (as he would do if he were the only person using the land), then those ranchers who are less altruistic and/or less foresightful will use up the resource instead, he will have sacrificed his own good for their sake, and the commons will still be degraded. To avoid thus becoming a sucker, he must use up as much of the resource as he can before somebody else does. And since every rancher will be in this same situation, the commons will quickly be degraded, to the detriment of all.
The second (alleged) tragedy of the commons situation (so called from its analogy to the foregoing situation) that Hardin is interested in is what he calls "the commons in breeding." He argues that the earth has a limited capacity of supporting humans, and that if we exceed some optimal population there will be severe negative consequences for human welfare overall. Furthermore, he thinks that if people are left free to breed as they will, human population will continue to increase, far above the optimum (it's not clear whether he thinks it will necessarily increase to the maximum sustainable level, as would consist with the analogy to the original tragedy of the commons).
This case is harder to make, though, since the individual parent doesn't necessarily gain anything by having more children. In fact, parents have to bear significant costs in time, energy, and money of raising additional children. Of course, they derive emotional rewards from having children, but it is doubtful whether one gets more emotional satisfaction from having ten children than from having two. So Hardin has to rely on another argument. He argues that if people are left free to breed as they will, then the people who want to have a lot of children will outbreed the responsible individuals who decide to have fewer. On the assumption that parents can pass on to their children the desire to have many children (if not genetically then by means of upbringing), society will soon become peopled by individuals who want large families, and the population will explode.
This argument is not entirely convincing either, because it depends on the assumption that children will inherit the desire for a large family from their parents. It's doubtful whether such a trait is genetic and still more doubtful whether parents who have large families teach their children to have large families also. Moreover, if Hardin's argument were correct, then the effect he anticipates ought to have happened by now, so that we should now be able to observe that almost everyone alive today desires a large family. Perhaps Hardin would reply that there are certain cultures in which large families are considered desirable. Still, it appears that this particular problem, if it is a genuine problem, nevertheless has nothing to do with the tragedy of the commons. The 'tragedy' in the commons situation is generated because each person has an incentive, as a rational egoist, to perform actions which are harmful to the group. But it is not the case that everyone has an egoistic reason for multiplying the number of his children. It is not as if people who now have only two children are rationally compelled to add more children to their family because others have large families. And as the world starts to become overpopulated and resources (e.g. real estate) become more scarce, there will be ever less incentive for people to have children, so the supposed population problem is not a tragedy-of-the-commons problem. The reason Hardin thinks it is is apparently that he assumes there is some inherent advantage in having a large family; thus, he wonders,
... how shall we deal with the family, the religion, the race, or the class (or indeed any distinguishable and cohesive group) that adopts overbreeding as a policy to secure its own aggrandizement? (1246)
But how overbreeding makes one better off is never made clear.
Now let's give an abstract definition of the 'tragedy of the commons' situation. We've been relying on the story about the ranchers with the common grazing area, and calling a 'tragedy of the commons problem' any problem that is analogous to the situation of the ranchers. But this is imprecise. Analogous in what way? What are the features of that case that define 'the tragedy of the commons' as a general type of problem, and that figure in a general argument in favor of social coercion?
Essentially, we have a number of agents (the fact that there are at least several individuals is important) each of whom has a certain course of action open to him. He benefits individually by performing that action, as compared against his refraining; however, every person is worse off if everyone undertakes this action than he would be if no one did it. Under these circumstances, the individuals caught in this situation are all forced into the course of action leading to the detriment of all. It is worth noting how this situation compares with and how it differs from the classic Prisoner's Dilemma. In the prisoner's dilemma, each of two agents is situated such that, if he does A, he will be better off, regardless of what the other agent does; however, both agents are worse off if they both do A than they would be if neither did A. Now the commons problem differs, first, in that many individuals are involved, and second, in that the individual's benefit from performing A is contingent on some other people's doing A. To explain the second point: In the commons problem, it may well be that each person can calculate that, if everyone else refrains from doing A, then he will be better off if he refrains from A also; but if someone else does A, then he will be better off if he does A. This is unlike the prisoner's dilemma, in which the individual agent benefits from doing A even if (in fact, especially if) the other agent refrains. But in the commons situation, it does not matter to me that if the other people cooperated, I would benefit by cooperating also. I still must defect, because if even one other person defects, I will be much worse off unless I defect as well.
Traditional game-theory idealizations (i.e., the assumption that every agent knows that all agents are rational egoists) may not explain this result, but there is little doubt that it is correct. Consider the situation of the ranchers, and assume there is a large number of ranchers using the land. Each of them can say to himself:
If my fellows all restrain themselves and only allow moderate grazing (x cows for each of us, say), then it will also be in my interests to be similarly responsible. For if I too put only x cows on the land, the commons will be preserved; if I try to put more cattle on the land, the commons will be degraded and eventually I will be worse off. However, if there is even one shortsighted person among us or one person who thinks he is entitled to more than the rest, then he will be able to quickly use up the grazing area, leaving it useless to the rest of us. In this case, I will be left with almost nothing, unless I too try to feed as many cows on the land as possible immediately. Furthermore, even if there is no shortsighted person among us, but there is someone who believes that others are shortsighted, the same result will follow. And since it's highly likely that one of these kinds of person is among us, it is best for me to go ahead and put as many cows on the land as I can, before someone else does.
This conclusion is reached the more certainly as there are more agents involved in the situation, since the number of people present increases the probability that there is some shortsighted individual, or else some individual who believes that others are shortsighted (or someone who believes that others believe that others are shortsighted, etc.)
We can now also see, incidentally, why one of the favorite solutions for the prisoner's dilemma won't work for the commons problem. It's now a commonplace to observe that if people play repeated 'prisoner's dilemma' games with each other, and they remember their partners' past behavior, the dominant strategy ('tit for tat') will generate cooperation. Iterating tragedy-of-the-commons situations, however, will not produce any such satisfying solution, because if there is even one person in the group who, for whatever reason, decides to always defect, then he wins out over all the cooperators, all the time. Thus, "always defect" is still the dominant strategy.
Here, then, we are faced with a problem. Supposing that commons situations exist in our society (the management of natural resources is the most likely example), what, if anything, can be done about them, to avert the disaster that ensues if every one acts as a rational egoist? Garrett Hardin suggests this solution: we can agree to establish some central authority which will force us all to cooperate. This is what he calls "mutual coercion, mutually agreed upon." Thus, our ill-fated ranchers might get together and pick some impartial party to be the 'police'; let's say they pick Bob. They give Bob their guns and tell him, "Now make sure none of us puts too many cattle on the land," and presumably they pay him some compensation for performing this job. Then they all go home feeling relieved and secure.
But another problem arises, which Hardin recognizes but does not have much of a solution for: Now that Bob is charged with watching the ranchers, who will watch Bob? If the ranchers can not be assured that Bob will do the job they have given him competently, and also that he will not otherwise exploit them, then they have not found much of a solution to their problem. Bob might disappoint the ranchers in various ways, once he is given power: (a) He might fail to protect the commons due to ignorance of what was required; he might not know how many cattle was too many to let use the land. (b) He might fail to protect the commons due to lack of interest; after all, if it is not his land, he may not care if it is degraded. (c) He might decide to exploit the commons himself; he might start raising his own animals on the land, and keep off the ranchers' cows. (d) He might demand exorbitant fees from the ranchers for his services. How, then, can the ranchers be sure that Bob will not attempt to do these things? We have assumed that Bob has been given a certain amount of power for purposes of policing the ranchers; this power is now what will make him difficult to control. Since he has sufficient power to force the ranchers to use the commons responsibly, there is reason to suspect that he will also have sufficient power to exploit the ranchers and their land.
Of course, this story is most interesting as a metaphor: if we as a society establish a coercive institution (in particular, a government) to force individuals to behave in ways that are beneficial to the group, how can we be sure that this institution will use its power to serve the interests of society and not rather merely to serve its own interests?
The first solution to this problem we might come up with would be this: if Bob does not do his job properly, the ranchers can get together and collectively oust him (although none of them individually would be able to stand up to Bob). Hopefully, the threat of this will keep Bob in line.
This plan, however, would seem to confront a problem analogous to the original tragedy of the commons. Each individual rancher would have to decide whether to join in the effort to overthrow Bob. By joining in, he incurs a certain risk to himself, but he also increases the chances that the rebellion will be successful. The problem is that the costs involved in his individual decision to help overthrow Bob will be borne by himself, whereas the benefits will redound to the group. So he will not join, preferring to stay at home and let the others take care of Bob; and, following the same reasoning, neither will anyone else, even though the whole group would be better off if everyone joined than if no one did.
This result is much clearer if we consider the possibility of overthrowing a government, for in that case it is still less likely that one individual's decision to join the rebellion will make the difference to whether it succeeds, but it is highly likely that it will result in that individual's death. We have, then, what we might call an inverted tragedy-of-the-commons problem ('inverted' because "benefit" and "harm" are interchanged): each person has an action available to him that would harm him individually but would benefit the group; and the logic of the commons predicts that no one will choose that action, even though the group as a whole is worse off in that case than they would be if everyone acted.
A second possible solution to our problem (in the context of the government of society) would be this: we have a democratic government. We can periodically elect leaders to tell us what to do, and we can vote out of office those who do not do their jobs well. Although this proposal makes for some improvement, it still introduces another form of the tragedy of the commons. Discovering which policies are desirable, which elected officials have performed well, and which candidates will perform well in the future, all requires extensive research. The individual who chooses thus to watch the government takes on all the costs in time and energy of his decision, while the benefits of his action are shared by the group. Under these circumstances, as the tragedy-of-the-commons logic predicts, very few people will expend their resources becoming informed about and involved in political matters. And, again, the individual who chooses so to expend his time has almost no chance in a large society of noticeably affecting public policy.
Here we see the explanation of a phenomenon that is commonly lamented in modern times: people in a democratic society usually don't vote, and when they do, they are usually ill-informed. This fact is often attributed to voter 'apathy', but it simply reflects the rationality of the voters, given the situation in which they are placed. Citizens realize that researching political issues and voting is a waste of their time. What we should actually find surprising is how many people do vote. (I suggest this is due to the fact that people do have a small amount of altruistic sentiment, and voting takes only a little time; it's also probably due to propaganda about the importance of voting.)
Perhaps the answer is that we just have to hope for our leaders to be responsible and benevolent. The question then becomes whether this is any more realistic than hoping for the ordinary citizens to be responsible and benevolent to begin with (in which case the tragedy of the commons problem would not have arisen). There is, on the contrary, reason to expect leaders of a coercive institution to be less altruistic than ordinary citizens. Those who choose as a career making rules that the rest of society are forced to obey, are likely to be those who value power and enjoy exercising power. Such individuals, I think, are less likely to be altruistic than the general run of men, and they will have a tendency to extend their own power whenever they can. Furthermore, the opportunities for benefitting oneself at the expense of the rest of society are certainly greater according as one has more power, so it is not clear that we have gained anything by placing power in the hands of a few individuals in order to prevent the others from selfishly harming society.
The solution to Garrett Hardin's original tragedy of the commons problem is fairly straightforward: the ranchers need a system of private property. If we can suppose the ranchers to be sufficiently coordinated and reasonable to be able to get together and agree to set up a government-like institution, then they ought to be coordinated and reasonable enough to be able instead to agree to divide up the commons into individually-owned parcels of land. Furthermore, they don't need even this degree of coordination if each rancher simply claims a plot of land if he is the first person to find and use it (in Lockean fashion). Each rancher then will have a selfish reason to maintain the quality of the grazing area because only so will he be able to continue to use his land in the future. The benefits of his adding another animal to his herd are still reaped by the individual rancher, but now the costs are also borne by him individually. And the ranchers are unlikely to try to graze cattle on someone else's land, or to try to take over each other's land, because, if they are rational, they will wish to maintain peace with their neighbors. So it appears that in this case there is no need for an external authority.
But this is not the end of the story, unfortunately. What about resources that can't be conveniently parcelled out, such as air and water? We can have a tragedy of the commons arise here, as Hardin points out, because individuals may have egoistic reasons to pollute the air or water that the public uses. If I release pollution into the air, I typically sustain the benefits (or averted costs) of my action, while the harm is distributed among society at large; but if everyone pollutes indiscriminately, the consequences are disastrous, so this is a genuine commons problem. And it is not realistic to suggest that we divide up the atmosphere into individually-owned parcels.
There are many possible instances of the commons problem, of which this is only one, so it is important to keep the abstract structure of the problem in mind if we wish to sketch a general solution to it. Since the problem arises when an individual has an act A open to him, the benefits of A will be borne by the agent, and the costs will be borne in large part by others, what we would like is to have a means whereby the costs, at least approximately, can be shifted back to the agent. This is attained in the original problem with the ranchers when the ranchers each own the land they are using; but that kind of solution is not always available. A more generally applicable solution would be this: each individual should be held liable (in court) for the costs that he imposes on others. For example, if someone releases noxious chemicals into the natural environment, then nearby people who are adversely affected could sue the polluter for damages.
A number of things need to be clarified about this proposed solution. First, since it relies on the court system, one might wonder how it really differs from Hardin's solution in which an external authority imposes regulations. As far as I understand it, Hardin's proposal (which is the traditional answer to the commons problem) is that the higher authority (the government) should simply decide directly who may do what -- e.g., who may produce pollution, of what kinds, and how much -- and then punish people for noncompliance. My more free market proposal has several advantages over this: (a) it demands less wisdom on the part of the authorities, since they don't have to know beforehand how much pollution should be allowed, and since they are not required to balance costs of pollution against benefits. (b) It allows little opportunity for abuse of power, since the authorities (the courts) in my proposal are only charged with deciding for or against a plaintiff in a given case and are not given broad powers of deciding how everyone should behave. (c) It makes use of market forces to determine the ideal amount of pollution -- polluters who can most cheaply reduce pollutants will be the ones encouraged to do so.
Second, it might be complained that there is often considerable difficulty identifying a particular individual as the cause of a specific harm. For instance, if I find my lake is more acidic than normal, presumably due to pollution, it may be difficult or impossible for me to identify the polluters who are responsible. Admittedly, this creates a problem, but it is also a problem for the alternative proposal: if we either can't tell who is producing pollution or can't determine what harms it causes, then it must be just as difficult for a regulative agency to decide how to regulate it as it is for a court to decide how to assign liability.
Finally, I have a second non-coercive solution to the commons problem to propose. It is that public opinion may serve to regulate people's, and companies', behavior. In general, if an individual acts immorally, or if he acts so as to harm his fellow men, then the other members of his community are likely to dislike him, and this will have negative consequences for him. For instance: Companies know that it is important to maintain good public relations. They also know that people are concerned about the natural environment. If some company becomes known for being an egregious polluter, then citizens can decide not to patronize it. Furthermore, environmental groups could publicize information about a company's environmental record. This solution does not seem to demand an unreasonable degree of altruism on the part of citizens, since deciding not to buy products from certain companies is generally easy and imposes very little cost on individual consumers, and since there already are numerous dedicated environmental groups.
This seems to me better than Hardin's coercive solution, because it does not create any new opportunities for individuals to profit at the expense of society, as the establishment of a coercive power inevitably does.
Hardin, Garrett. "The Tragedy of the Commons", Science, Dec. 13, 1968 (pp. 1243-8)
Anderson, Terry and Donald Leal. Free Market Environmentalism (San Francisco: Pacific Research Institute, 1991).