This is the second part of a comprehensive essay on the Rawlsian view of CSR (corporate social responsibility) seen as a multistakehodler model of corporate governance. I describe the repeated game among the firms and its stakeholders so that several types of reputations, based on the full or less than full respect of the CSR model may be developed. Here the Rawlsian social contract performs its main role: that is, the ex ante impartial selection of a unique equilibrium amongst the many possible in the repeated trust game involving the firms and its stakeholders. In this context it allows impartially selecting just one fair reputation equilibrium amongst the many possible. Elaborating on Binmore’s Natural Justice (2005) (but see also Binmore, 1987, 1991, 1994 and 1998) and it reevaluation of John Rawls egalitarian and maximin principle of justice within a game theoretical perspective, this task is accomplished from the ex ante (under the ‘veil of ignorance’) point of view, but in a way that allows to find out a unique course of action that satisfies the requirement of incentive compatibility (i.e. a Nash equilibrium). According to Binmore, and contrary to the belief that Rawls’ view was utopian, it is shown that the maximin principle provides the best account of the social contract under the assumption that in a ‘state of nature’ any agreement on principles for institutions must be self-sustainable. Such an unconventional result has overarching implications also for the constitutional contract on the firm’s governance and control structures. This is a theory to make sense of the idea of extended fiduciary duties put forward in previous works (Sacconi, 1997, 2000, 2006a,b, 2007, 2010a). Its main point was that the stakeholders’ constitutional agreement (seen as the rational solution of an original bargaining game) will complement the efficient control structure with further social responsibilities toward non-controlling stakeholders. However, when a constitutional bargaining situation is considered such that the only feasible constitutions are allocations of exclusive property and control rights, a strong imbalance of bargaining power is inevitable, so that asymmetry in the final surplus distribution will reflect the asymmetry of decision rights. Then, an outcome corresponding to the arrangement of rights that immunizes non-controlling stakeholders against abuse of authority, may not belong in the equilibrium space of the constitutional choice game. Many scholars of corporate governance accustomed to accepting second-best solutions would then be ready to give up fairness and extended fiduciary duties in order to achieve nothing more than the most efficient constitution of the firm. Remarkably enough, application of the Rawls-Binmore theory to the social contract on corporate governance structures yields quite the opposite suggestion. In order to be consistent with the requirement of self sustainability, the impartial agreement must select the constitution with the best egalitarian solution among all the alternative feasible constitutions. Pareto dominance, as a principle of unanimous agreement, is therefore to be applied only to the comparison of feasible egalitarian solutions under alternative constitutions. What is most important here is that this result follows straightforwardly from the requirement that the social contract should select an outcome belonging to the set of (impartial) equilibria, i.e. a self-sustaining institution. Quite unconventionally, here fairness precedes both efficiency and welfare maximization (contrary to Kaplow and Shavell), and it also precedes aggregate transaction costs minimization (against Hansmann 1988, 1996). Even libertarians like Hayek’s followers - who typically believe that rules of behavior should spontaneously emerge from endogenous motivations respecting free choice – will have to concede that under the simple ethical constraint of impartiality egalitarianism is a natural consequence of the self sustainability of institutions in the domain of corporate governance
