The Moral Bargain…”The Wind Cries Mary”

Welcome to another installment wherein I attempt to tear assunder libertarianism and social justice. Two things the likes of LoOG, Let the wind between howl, lonely and wild, and BHL,The Moral Status Quo, are intent on marrying.

Let me begin by noting that BHL’s allusion to a “frozen conceptual sea” as a critical starting point nonetheless fails to extricate the discussion from the glaciers. BHL’s definition of “classical liberal” and “libertarian” are standard fare. The introduction of the term “High Liberal” offers little clarity and fails to crack the ice.

I take a different approach, one that is rooted in contractarianism and not economics per se. I define Liberalism thusly:

Liberalism: a rational normative social contractarian theory that views morality and justice to be a product of convention and agreement.

I would contend that this is universal definition, one that holds even today among the various factions. Divergences occur, however, around the different conceptions of what it means to act rationally. This is the basis of the fault lines within liberalism. And these fault lines have been there from the start. Examples:

(a) For Hobbes, the rational agent will cede it’s liberty to obey a sovereign.
(b) For Locke, the rational agent will only cede the right of retribution to the sovereign as an enforcer.
(c) For Kant, the rational agent will cede authority to a universal moral principle as means to ensure personal autonomy.

From a contractarian perspective, a primary distinction between “enlightenment liberalism” and modern liberalism is the shift from legitimacy of political authority to that of justice vis a vis the social contract. For the latter, we can certainly point to Rawls. For Rawls, the social contract is less about legitimizing political authority per se and more about justifying political and socio/economic arrangements.

There are three basic categories of justice that can be distinguished:

(i) Justice of Impartiality: equality of the moral agent
(ii) Justice of Reciprocity: proportionality, fair exchange
(iii) Justice of Mutual Advantage: outcome of a bargain

Game Theory, a framework for analyzing strategic decisions under uncertainty, provides a model for contractarianism that treats it as a problem in a type of cooperative game. Bargaining solutions to cooperative games can perhaps shed some light on agent rationality.


Rawlsian Justice is primarily concerned with (i). Rawls can be recast as fairness as a foundation for a “rules universalism.” Rawls’ “Difference Principle” is not the result of a bargaining outcome between rational agents–that is, an outcome rational agents should choose. Rather it is a decision rule an unbiased(impartial) moral agent should follow. Ralwsian Justice is not a solution to a coordination problem in a cooperative game, but rather it is treated as an initial value problem of the unbiased moral agent. Rawls resolution to the initial value problem relies on the hypothetical construct of the “Veil of Ignorance.” Rawls’ version of VOI, which is usually described as the “thick version,” simply means that unbiased agents will reason the same when it comes to “fairness.” Under a “thick veil,” they will maximize the minimum position. So any problem of universal choice reduces to the preferences of an unbiased individual agent.

However, I think Rawlsian Justice fails the empirical test with it’s reliance on “fairness” as something that can be universalized across individual choice under the same initial condition. Experimental Games suggests that “fairness,” in terms of any universality, is rooted much more in bargaining to prevent/punish “gaming advantage of the rules,” cheating and free riding. “Fairness” is much more a product of a bargaining outcome and much less a hypothetical something grounded in an “originalist position.”

Politically, the “cultural war” presents damning evidence against Rawls. We can model the “cultural war” as conflict between cultural identity groups fighting over redistribution. The basis of the conflict are “redistribution rules” that are viewed as allowing cultural identity to be a source of unfair advantage. A Distributive justice rooted in Fairness/Unfairness that are products of political conflict results in something quite different than a “Difference Principle” and a basic set of primary liberties(primary goods). Instead, you end up with a basic set of “primary obligations” that must be met in order to avoid the “min-min” principle, which is minimizing the minimum.

Rawls suffers from a “Redistribution Problem.” This cannot be “normalized” away vis a vis hypothetical constructs that have no basis in reality. It cannot avoid the rigors of cooperative games. It doesn’t hold up.

Libertarianism and the Justice of Mutual Advantage

Historically proper, Libertarianism should be understood as a radical subset of liberalism that rejected any contractarian basis for political authority. This is the unifying principle within libertarianism. In terms of Economics, however, there has been nothing but bitter dispute within this tradition. This is one reason why it is nonsense to try to use libertarianism as some economic justice bridge between “classical liberalism” and “modern liberalism.” There is no bridge there.

Whereas Rawls avoids bargaining, libertarian contractarianism, due to cooperation creating “additional surplus” that can incentivize equalizing constraints(concessions) on “agent rational self interest,” holds up potentially quite well under cooperative game theory. The Kalai-Smorodinsky bargaining solution provides a rational basis for a Justice of Mutual Advantage. David Gauthier used this to argue for “Morals by Agreement.” However, Gauthier uses a “minimax relative concession” constraint in a one-shot non-cooperative game to arrive at the same outcome as the Kalai-Smorodinsky bargaining solution. Gauthier’s Moral Contractarianism does not rely on any assumptions or hypothetical constructs of impartiality or empathy; the only requirement is the rational agent distinguish between short-term interest and long-term interest and realize other rational agents will modify their behavior based on expectations of how others will behave–in other words, they understand the concept of games.

Gauthier’s rational agent will behave very differently than Hobbes’ Rational Agent. The Gauthier Rational Agent Paradigm:

(i) Cooperate with cooperators
(ii) defect from defectors
(iii) cooperators will play a strategy to minimize: [u(m) – u(C)]/[u(m) – u(d)]
(iv) u(b) >= u(d)

u=given utility function for each agent
u(m)=max utility for the agent in a bargain that would lead to other agents to defect. The breaking point(or, the ‘Donald Trump Art of the Deal” max advantage for the given agent)
u(C)=utility of a concession C
u(d)=utility of the defection state, no deal…
u(b)=utility of the bargaining equilibrium

(iv) is usually referred to as the “Lockean Proviso,” which, in this context, means that a moral bargain should not result in any agent being worse off than the defection state. In other words, cooperation=Win/Win. It can also be interpreted as the libertarian principle.

JMA, in essence, is a less theory of justice and more a theory concerning the conditions under which moral constraints will be effective. Interestingly, Gauthier himself is not self-described libertarian. Jan Narveson is usually credited with advancing a radical libertarian interpretation of Gauthier’s work.

Problems with Justice of Mutual Advantage

A Conceptual Problem
The immediate criticism of Justice of Mutual Advantage is that it is too exclusive. We can denote this as the “Vulnerability Problem” which pertains to those who are unable to contribute or contract. This usually relates to “Children.” One possible resolution to this problem is note that experimental studies suggest that impartial moral judgements are emergent properties of JMA. This echoes Adam Smith’s Theory of Moral Sentiments(empathy as the product of the Invisible Hand). So JMA may not be able to directly resolve the Vulnerability problem but it potentially creates a derivative moral foundation(Justice of Impartiality) that can extend the benefits of justice to all its members.

The Empirical Problem
The obvious empirical problem is that the social order is not a population of equal, optimal, and tolerant cooperators. This, in one sense, can be explained by the fact that bargaining equilibriums in cooperative games are generally not unique. Constrained Maximization is not a unique equilibrium. And, in general, in the modeling of complex social eco-systems, rational choice should be supplemented with biology(evolutionary game theory). In a more expansive model, agents are not only rational strategic reciprocators, but are also sensitive to the perceived altruism preferences of other Agents. Reciprocity is key to stabilizing cooperation, but under evolutionary competition, altruistic cooperators may be more evolutionary fit than Gauthier’s constrained maximizers. Competition between altruistic cooperators can result in maximizing the number of altruistic preferences, which can lead to a type of race to the “moral high ground.” Reciprocal mechanisms implemented can result in exploitation of the lesser fit altruistic cooperators(those who are slow to the moral race) by the most fit altruistic cooperators.

In short, a “Punishment Problem” potentially emerges in evolutionary game theory for a model that relies on reciprocity as the basis of cooperation.


Libertarian justice, understood in terms of contractarianism, is a theory of rational moral constraints. In modern game theory, these constraints result in a equilibrium that is equivalent to the robust Kalai-Smorodinsky bargaining solution. This is one of the more important developments in libertarian theory the past 25 years, in that it ties the libertarian principle to a boundary constraint on non-cooperative games. In this framework, Justice Categories like Impartiality are treated as derivative emergent properties that do not violate the boundary constraints.

Libertarian Justice modeled as moral constraints is quite different than Rawls which treats justice as an initial value problem. Rawls offers no rational constraints on individual choice outside the initial state, which is why he intractably suffers from a “Redistribution Problem.” At this point, I hope I have adequately made the case that Libertarianism as moral boundary constraints on initial value Social Justice principles is akin to squaring the circle.

To me, one of the more relevant problems of Libertarianism would be the “Punishment Problem,” namely moral constraints on punishment in the context of reciprocity being a necessary component of stable cooperative equilibrium. This is a biological problem…

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s