Two Cheers for Mutual Advantage
In this post I want to explore a dimension of politics that seems to have been wrongly marginalized in discussions of political conflict and polarization. This is the seemingly simple, innocent idea of cooperation for mutual advantage. No plausible moral or political theory rejects the idea of mutual advantage entirely, but a dominant theme in recent political thought emphasizes the limits of mutual advantage rather than its extent. What needs greater emphasis at the present moment in my view is the cooperative rather than conflictual dimension of political choice.
The most influential conception of mutual advantage is arguably the Pareto principle in economics. The Pareto principle is a generalization of mutual advantage from the two-person to the n-person case. It says (roughly) that a distribution of goods, A, is Pareto-superior to a distribution B, just in case no person is worse-off in A than in B, and at least one person is better-off in A than in B. On this view, mutually advantageous actions or exchanges are those that create Pareto-superior distributions. A distribution is Pareto-optimal just in case there is no (feasible) distribution that is Pareto superior to it. In other words, Pareto optimality means there is no distribution that makes at least one person better off, and no one worse off, than the current distribution.
In the context of welfare economics, the Pareto principle is understood as a condition on a collective choice rule (CCR). A CCR is a function that takes as its domain a vector of individual orderings over different distributions and spits out a particular “social ordering” over those distributions. In essence, the Pareto condition on CCRs says that if everyone prefers distribution A to B, then A is socially preferred to B. So described, the Pareto principle sounds unobjectionable.
However, the Pareto principle has limitations. One limitation is that the Pareto principle is indeterminate. Many distributions will not be Pareto-comparable to one another, including those distributions that are all Pareto-optimal. While it is possible to slightly amend the principle to treat all Pareto-optimal distributions as equally good, and therefore comparable, this raises an ethical objection in the form of insensitivity to distribution. The Pareto principle does not distinguish between equal and unequal distributions: a distribution in which Warren Buffett gives a dollar to a homeless man is just as good as one where the homeless man gives a dollar to Warren Buffett. For these reasons, most writers hold that the Pareto principle can be at most a necessary condition for a satisfactory CCR rather than a sufficient condition.
One interpretation of the indeterminacy in the Pareto principle emphasizes its insensitivity to conflictual or zero-sum elements of political choice. Many political choices do involve making some better off at the expense of others. However, there is a more positive interpretation as well. This is the connection between the Pareto principle and unanimity as a decision rule. Intuitively, many of us think of unanimity as the result of convergence on a single outcome in deliberation. Unanimity selects the one outcome everyone can accept. Depending on the status quo, however, there are often many feasible outcomes that can make everyone better off. The difficulty then becomes how to choose among these alternatives.
There is a famous example in the history of philosophy from David Hume’s Treatise of Human Nature that illustrates this connection between indeterminacy and unanimity, and which is often cited in discussions of collective action problems and free riding:
Two neighbours may agree to drain a meadow, which they possess in common; because it is easy for them to know each others mind; and each must perceive, that the immediate consequence of his failing in his part, is, the abandoning the whole project. But it is very difficult, and indeed impossible, that a thousand persons should agree in any such action; it being difficult for them to concert so complicated a design, and still more difficult for them to execute it; while each seeks a pretext to free himself of the trouble and expen[s]e, and would lay the whole burden on others.
Hume’s example of draining a meadow is usually read as an example of a free-rider problem in the large-number case. In the case of two persons, each party’s contribution is necessary to securing the good, so both parties have an incentive to contribute to draining the meadow. But when a thousand people are involved, all of whom can enjoy the benefits of draining the meadow without contributing, none has an incentive to contribute individually. The difficulty Hume is actually pointing out in the large-number example, however, is not a free-rider problem but a bargaining problem. This is because there would an incentive to cooperate if draining the meadow required all 1,000 people to contribute. Not only would everyone have an individual incentive to cooperate, but there would be an incentive to make sure others were cooperating too (“Keep an eye on Jim… he might be the weak link here”). Instead, a far lower number of people is strictly necessary to drain the meadow (maybe only two!). The advantage is economies of scale — that spreading the cost over a 1,000 people means that the cost to each could be very low.
The problem is that there are an enormous number of ways to divide the burden of draining the meadow between 1,000 people, especially when we introduce something like side payments. There is the psychological and practical difficulty of simply coordinating such a large number of people on a single proposal. But also, if we require all one thousand to agree to any proposal using a unanimity rule, there is a huge incentive for strategic behavior. If one person can block a proposal, then each person can try to hold out and shift the burden of draining the meadow on to others (“each seeks a pretext to free himself of the trouble and expense…”).
Free-rider problems and bargaining problems (as I’m using the expressions) are both barriers to collective action, but they are different kinds of problems. Free-rider problems arise from the difference between the value to me of collective action and the total value of collective action to all in the group. Bargaining problems, on the other hand, are problems in the distribution of benefits from collective action, not the problem of whether or not collective action will occur.
Why does this distinction matter? This brings me back to the connection between unanimity, Pareto, and indeterminacy. Bargaining problems take place on the assumption that parties are trying to reach a mutually advantageous (Pareto efficient) agreement relative to the status quo. Opportunities for bargaining arise from the existence of many different mutually advantageous alternatives. Skilled negotiators know the importance of excluding certain alternatives — this is the intuition behind “burning your bridges” that Thomas Schelling explored at length in The Strategy of Conflict. Put another way, the problem with Pareto efficiency is how to cooperate for mutual advantage, not whether we should. There are so many roads to mutual advantage that it is easy to pass over in a rush to confront the question of who gets more.
The Swedish economist Knut Wicksell gave one of the clearest statements of the connection between unanimity and mutual advantage in his 1896 essay, “A New Principle of Just Taxation.” The central thesis of Wicksell’s essay, I think, is well-captured in the following quotation:
Provided some [public] expenditure holds out any prospect at all of creating utility exceeding cost, it will always be theoretically possible, and approximately so in practice, to find a distribution of costs such that all parties regard the expenditure as beneficial and may therefore approve it unanimously. (pp. 89–90 in Musgrave and Peacocke, Classics in the Theory of Public Finance, St. Martin’s Press: NYC, 1958)
In Wicksell’s view, the thesis applied to all genuinely public goods, since if there were no price (in the form of a tax share) that some person would be willing to pay, the good or service would be properly understood as a private enjoyment rather than a matter of public concern. The way to ensure this criterion of genuine publicity would be met would be to tie any proposed public expenditure to a proposal for raising revenue. The problem in public finance was how to find a distribution of costs that would satisfy mutual advantage and, in principle, receive unanimous approval. This problem is by no means trivial (for example, people might misrepresent their willingness to pay). But Wicksell’s point was that if each citizen could be taxed in accordance with their willingness to pay for a level of public service, each citizen could voluntarily accept the level of expenditure and their share of the cost together. This would be the closest political society could approximate a voluntary association.
Might there be cost distributions that would make some people worse off than the status quo and therefore fail the Pareto test? Sure, but those people would block any such proposal that made them worse off. And if negotiations broke down over such an extreme proposal, the real failure would be the failure to reach any mutually advantageous agreement at all.
Last month talks in the US broke down over a new COVID-19 stimulus bill. The sticking point was the size of the stimulus, and the lack of unified leadership within the GOP didn’t help matters. You cannot negotiate with someone’s agent if their agent (Mnuchin) can’t bind the principal (Trump), and when the principals (Trump & McConnell) aren’t on the same page. What seems to have been lost in the coverage is that the parties all agreed that some bill needed to pass and that inaction would be fatal. Part of the problem is language: “giving in,” “capitulate,” “cave” are all zero-sum phrases that evoke winners and losers. What “giving in” actually means in this context is that your side doesn’t get everything you want, not that you lose what you have.
Another way to undermine the grim logic of zero-sum conflict is to take clearer inventory of the cooperative dimensions to social life. One of the reasons cooperation in everyday life is easy to overlook is that the stakes are often low. For whatever reason (possibly the influence of Christianity or Kant?) there is an idea that altruistic or moral action worth its name has to involve some huge sacrifice on the doer’s part. Because these situations are rare, it is easy to conclude that most people aren’t altruistic. What this misses are cases like drivers letting someone merge on the highway. As far as rational self-interest goes, it is a mystery why any driver ever lets someone else get in. Let the car behind me do it! It’s not like someone will ticket me if I don’t. While less pleasant encounters in traffic come to mind more easily, these interactions happen more often than we realize. Communication is another cooperative practice that is easy to overlook. We again fixate on high-stakes cases where people lie or deceive. This makes it easy to overlook lower-stakes opportunities for misrepresentation or free-riding that people forego. I think of the student who asks the “dumb” question that is in fact on everyone’s mind in class or someone who stanches rather than spreads a rumor.
The deep problem here is that human societies are a mixture of cooperation and conflict. Conflict means we cannot be naïve, but cooperation means our fellows aren’t enemies. Most circumstances for conflict — market competition, politics, fighting — take place against some background norms of cooperation. We rarely accept conflict for its own sake and more often we accept it because it secures someone cooperative benefit. The philosopher Agnes Callard makes this point nicely about fighting. Fighting over a prize is appropriate when there is no independent standard to determine who should get the prize and each claimant is uncertain whether she deserves the prize. In this case, fighting can be the best way to reveal the party with the superior claim. Callard’s point is that fighting isn’t only about beating the other party. It’s a rational response to a genuine gap in our knowledge and especially our self-knowledge. You only know what you’re “made of,” so to speak, when you fight. In this way, even fighting is a zero-sum interaction in the context of wider, variable-sum pattern of interaction. Both participants, as well as concerned third parties, can be better off than they were.