Distributive Justice with Production and the Social Contract. An Experimental study

14 April 2016
Versione stampabile

2 PM 
Room 3E, Department of Economics and Management


  • Giacomo Degli Antoni, University of Parma


The aim of this paper is to provide experimental evidence of the influence that explicit agreement under the veil of ignorance may have on individuals’ conception of justice and its implementation in a context of production of a common output to be divided among the individuals themselves. As we will explain, the evidence we are seeking should support or undermine one crucial supposition of the social contract argument as applied to distributive justice in general and business ethics in particular.
Drawing on the experimental literature on distributive justice, we put some theoretical assumptions of the contractarian argument to an empirical test through an experiment in which individuals play for real money in a laboratory situation that captures, in a simplified way, the idea of an agreement behind a veil of ignorance, and then the implementation of the agreed distributive rule(s) in a context of division of a real production outcome and absence of coercive authority. One crucial characteristic of our experiment is that subjects are given asymmetrical endowments, for which they are not responsible. At the same time, with their work they can generate unequal distributions of money. 
Do the subjects involved in this interaction distinguish between the two types of inequality? Do they try to reduce the arbitrary one, while accepting the one generated through work? Do they elaborate other distributive criteria? Do their choice ex-ante, when they are behind the veil, differ from their choice ex-post once the veil is lifted and they know the outcome of the production phase? 
The experiment is designed to find answers to these as well as to other relevant questions.
Our experiment is relevant for two stances of literature. First, it contrasts with experimental literature on distributive justice as evolved so far. On the one side, studies inspired by Rawls’s principles of justice, in fact focus almost exclusively either on the relevance of the difference principle (Brickman, 1977; Yaari and Bar-Hille, 1984; Frohlich et al. 1987; Frohlich and Oppenheimer 1990 and 1992; Bold and Park, 1991; Lissowski, Tyszka, Okrasa, 1991; Jackson and Hill; 1995; Michelbach et al. 2003; Cruz-Doña and Martina, 2000) or on the effects of the choice behind the veil of ignorance on preferences for redistributions (Anderson and Lyttkens; Straub et al. 2005; Herne and Suojanen, 2004; Herne and Marde, 2008; Schildberg-Hörisch, 2010; Durante, Putterman and van der Weele, 2014). None (except Faillo et al. 2015) take into consideration Rawls’ concept of “sense of justice” and its role in the solution of the problem of the ex-post implementation (compliance) of the principle chosen behind the veil. In the typical experiment, once the principle is chosen, it is automatically implemented. Considering the subset of  studies that explicitly implement the choice among alternative principles, participants usually choose individually and they are confronted with hypothetical scenarios (Scott et al . 2001; Konow, 1996, 2001, 2003). Only Frohlich et al. (1987) and Faillo et al. (2015) implements agreements among participants as a way of choosing the principles. In our experiment we make some steps onward with respect to this literature by taking into serious consideration the contribution of the contractarian tradition on distribution justice by i) introducing an explicit (non-binding) agreement behind a veil of ignorance – considering two alternative deliberative procedures;  ii) introducing a production phase in which the outcome depends both on luck and on individual effort, iii) studying the determinants of individuals’ willingness to agree, ex-post, with the agreement principle. 
Secondly, our study gives more substance to social contract approach to business ethics. By this term we intend broadly the so-called Social Contract tradition in Business Ethics (Wempe 2005), as represented eminently by Donaldson (1982, 1984), Donaldson and Dunfee (1999), Phillips (2003), Evan and Freeman (1988), Freeman (1994) and recently criticized by Mansell (2013). More specifically, however, we draw on Sacconi’s view of the firm as an institution that can be normatively reconstructed as the result of a constitutional agreement among its essential stakeholders (Sacconi 2000, 2006, 2010a,b). This view is in line with Freeman’s (1989) stakeholder perspective (refined in Freeman 1988, Evan and Freeman 1993; Freeman 2011), and with Aoki’s (1984, 2010) cooperative theory of the firm. For its theoretical normative core, it rests on Binmore’s (2005) re-casting of Rawlsian contractarianism (Rawls 1971), and Buchanan (1975) logical reconstruction of constitutional and post constitutional contracts. The basic idea is that the normative basis of the firm can be conceived as the result of a unanimous hypothetical agreement among all the firm’s constitutive stakeholders. The content of the agreement would be the set of principles and governance structure that define a mutually beneficial cooperative venture, within the rules of a market economy. 
In this experimental work, however, we enquire whether real life subjects, participating in a situation where the relevant intuitions of justice concerning the division of the joint output of a productive activity are involved, would actually follow principles consistent with the social contract perspective in business ethics as sketched above. Obviously, we do not claim that our experiment represents all the complexities of the firm’s productive activity, and its related governance structure. Only a stylized representation of some features of it can be subjected to laboratory experimentation. Nevertheless, with respect to the theoretical literature on the social contract of the firm, our experimental study adds an element of realism. We take a concrete situation wherein production means are not initially distributed according to any principle but chance, and parties in a production activity bringing about a common output have to make a choice how to divide it amongst the producers. Hence, we enquire empirically whether these real life agents, being put under an experimental simulation of the Rawlsian “veil of ignorance” that guarantees they assess impartially principles of division, would converge - not just in principle but also in practice - toward the distributive principle recommended by a social contract. This is a principle of distribution that, according the liberal egalitarian view, would rectify the effects of an initial arbitrary allocation of rights over production means, and would disentangle such effects from those attributable only to personal contribution and desert. One could say that seldom real economic environments or the governance structures of most firms de facto satisfy (and by sure do not satisfy by any logical necessity) ideal conditions of justice. Therefore our enquiry seems relevant to give substance to contractarian business ethics in seeing whether real life agents would uphold contractarian principles of justice in such non ideal productive contexts, so that they would be ready to rectify distribution outcomes after having entered the experiment of the ‘veil of ignorance’. 
Our analysis presents two main results. First, we find that liberal egalitarianism performs better than any alternative rule in our experiment. This preference however emerges from the subjects’ choices only when they go through a decision process under a ‘veil of ignorance’. In other words, subjects prefer liberal egalitarianism only when (i) they take an impartial and impersonal standpoint by subtracting information about their personal position in the initial endowments allocation; (ii)  at the same time they are aware of the necessity of reaching an agreement over principles. 
The second main result of this study is that subjects not just agree on Liberal egalitarianism behind the veil of ignorance, but also behave according to it in the ex post decision, when it is not at stake just the expression of judgments on principles but the actual distribution of an output sum. And, moreover, this happens in a contexts where they know that each of them has ½ probability to find herself in the position of a dictator - who may implement unilaterally her preferred division rule, included appropriating all the produced output (this is the typical nature of a ‘dictator with taking’ game).
These results are in line with both the theory of the ‘sense of justice’ and ‘conformist preferences’ in which an ex-ante agreement affects causally the formation of attitudes and preferences favorable to respecting fair principles agreed under a veil of ignorance. As a matter of fact, an high level of compliance is displayed in general for all the agreed principles. But we emphasize this result with respect to principles like liberal egalitarianism (or also simple egalitarianism), since in this case deciding to act upon the commitment of sharing ex-post the output of the unequal distribution of endowments is motivationally more demanding.
Summing up, moral reasoning behind the veil of ignorance concludes mostly to an agreement on the liberal egalitarian rule, which is a committment to redress unequal splitting of endowments ex post. Such a commitment provides a basis for an intention to act that may translate into a desire. So that agent would act not on a desire-based intention but on a commitment-based intention that may engender a desire (a reason-to-act-based desire). At the same time the committment-based intentional explanation constitutes a model for understanding other agents’ behavior. 

Keywords: Distributive justice, social contract, fairness, dictator game.
JEL Classification:C72,  C91, D02, D63