Hostname: page-component-cd9895bd7-lnqnp Total loading time: 0 Render date: 2024-12-28T01:13:27.803Z Has data issue: false hasContentIssue false
  • English
  • Français

Laboratory Results on Rawls's Distributive Justice

Published online by Cambridge University Press:  27 January 2009

Norman Frohlich ,
Joe A. Oppenheimer  and
Cheryl L. Eavey
Get access Rights & Permissions [Opens in a new window]

Extract

The behavioural underpinnings of Rawls's notion of distributive justice as outlined in A Theory of Justice are tested in experimental contexts. Under conditions approximating Rawls's ‘original position’ (including the appropriate agenda, a ‘veil of ignorance’ and a choice rule designed to capture his main theoretical constraints), we test his ‘predictions’ that individuals would reach a unanimous consensus on a principle of distributive justice and would select the difference principle: a principle that maximizes the welfare of the worst-off individual in the society. This view is contrasted with our belief, that any general concern for fairness (or distributive justice) will take a different form: one that both attempts to take into account several values and pays attention to cardinal rather than ordinal measures of utility. Our results strongly indicate that individuals are capable of reaching consensus but that they choose what Rawls has called an ‘intuitionistic’ principle which attempts to take into account not only the position of the worst-off individual but the potential expected gain for the rest of society. The overwhelmingly preferred principle is maximizing the average income with a floor constraint.

Type
Articles
Information
British Journal of Political Science , Volume 17 , Issue 1 , January 1987 , pp. 1 - 21
Copyright
Copyright © Cambridge University Press 1987

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 Rawls, John, A Theory of Justice (Cambridge, Mass.: Harvard University Press, 1971).Google Scholar

2 Nozick has, with a certain anathema, referred to this as ‘patterning’. Nozick, Robert, Anarchy, State and Utopia (New York: Basic Books, 1974), pp. 155–60.Google Scholar This Rawlsian conception of justice is contrasted with his own which is based on ‘entitlements’.

3 For a presentation and review of some findings about actual preferences and attitudes individuals have regarding income distribution and distributive justice, see Hochschild, Jennifer L., What's Fair: American Beliefs About Distributive Justice (Cambridge, Mass.: Harvard University Press, 1981)Google Scholar and Soltan, Karol E., ‘Empirical Studies of Distributive Justice’, Ethics, XCII (1982), 673–91.CrossRefGoogle Scholar

4 This notion, similar to the notions of fair division in game theory, is obviously a great simplification. For example, Nozick argues that individuals are not just dividing up a purely common property upon which they all make a claim. Rather, they are dividing a good to which they contribute differentially and in which they have property rights. Preliminary reports on experiments which consider this issue and the related issue of the stability of any distributive principle are contained in Frohlich, N. and Oppenheimer, J., ‘A Test of the Stability of a Political Conception of Justice in Experimental Groups’Google Scholar, a paper presented at the annual meeting of the American Political Science Association, Washington, D.C., 1986.

5 Although Rawls is ambiguous in a number of areas, he is absolutely clear in his acceptance of the contractarian view of justice.

6 Rawls, , A Theory of Justice, p. 17.Google Scholar

7 Rawls, , A Theory of Justice, p. 19.Google Scholar

8 This requires that there is an a priori agreement about the final nature of these considered convictions.

9 The constraints of the original position together with the psychology of individual rationality are then used to derive both the principles governing the distribution of wealth and income and the set of individual rights which inhere in the constitution of Rawls's just society. Although the moral status of these latter rights of individuals takes precedence over the principles of distributive justice, here we consider only the problem of choosing a principle of distributive justice.

10 Rawls, , A Theory of Justice, p. 64.Google Scholar Note that there is an obvious tension, which Rawls does not address, between this requirement and the condition of ignorance.

11 Rawls does not use the term ‘welfare’ but prefers to use the term ‘primary goods’. For convenience, we will use the terms ‘welfare’, ‘resources’, ‘primary goods’ and ‘income’ interchangeably.

12 Miller, Gary J. and Oppenheimer, Joe A., ‘Universalism in Experimental Committees’, American Political Science Review, LXXVI (1982), 561–75.CrossRefGoogle ScholarFrohlich, Norman and Oppenheimer, Joe A. with Boschman, Irvin and Bond, Pat, ‘Beyond Economic Man’, Journal of Conflict Resolution, XXVIII (1984), 3–24.CrossRefGoogle Scholar

13 Rawls, , A Theory of Justice, Section 7 and 49.Google Scholar

14 Rawls, , A Theory of Justice, pp. 154–5.Google Scholar

15 With a fixed economic product, this would imply an egalitarian distribution. But Rawls does not seem interested in justifying strict equality. Rather, he is concerned with distinguishing between justifiable and non-justifiable inequalities. Under many conceivable conditions the lot of the worstoff individual might be improved by incentives which generated not only inequality but also more economic product to be distributed. Excellent discussions of the theoretical relationship between inequality and the size of a society's overall GNP – especially as they affect fairness – may be found in Sen, A. K., On Economic Inequality (New York: Wiley, 1973).CrossRefGoogle Scholar Coleman argues for the desireability of Rawls's attempt to distinguish between justifiable and non-justifiable inequalities (Coleman, James S., ‘Review Essay: Inequality. Sociology, and Moral Philosophy’, American Journal of Sociology, LXXX (1974), 739–63).CrossRefGoogle Scholar

16 Indeed, one could argue that (political) ethics resembles the empirical sciences more than it does pure logic and mathematics, inasmuch as it has an empirical element. Thus our interest here is not in what some have called the geometry of morals, bul rather in the physics of morals.

17 The conception here is similar to a notion of robustness regarding scientific theories. Bridge concepts such as friction or air resistance serve to permit measurement of the approximation of ideal constructs such as ‘vacuum’. As the ideal is approached, the predictions of hypotheses which are defined for that ideal must increase in accuracy or the theory becomes uninteresting. Some theories are robust in only some dimensions (for example, in neo-classical market economic theory difficulties arise as we introduce certain violations of the idealizations). Failure to exhibit this sort of ‘global robustness’ would constitute a vulnerability or brittleness which would deprive the theory of a considerable degree of its attractiveness. We are endebted to Norman Schofield for these helpful insights.

18 Individuals often failed the first time, but subsequent failures were not frequent. We helped those with difficulties by answering questions about the principles they showed evidence of misunderstanding in the test.

19 This is done by having the subjects first see a set of income distributions, then choosing a principle of justice to select one of the distributions as the basis for their pay-off, and finally being randomly assigned to an income class by reaching into a bag and pulling out a chit. Each chit not only shows a class assignment and the associated pay-off, but also specifies how much they would have received, given this class membership, for each of their possible choices of principle. This procedure is repeated four times. Each time, chits are kept by the subjects to allow them to reflect on the consequences of their choices. Judging from the discussions in the experiments this had considerable pedagogic impact.

20 To make it credible the $40 amount was actually written on each student's pay-off sheet – on which the amounts the subject had earned in the first part of the experiment were recorded.

21 Tversky, A. and Kahnemann, D., ‘The Framing of Decisions and the Psychology of Choice’, Science, CCXXI (1981), 453–8.CrossRefGoogle Scholar

22 Specifically, the index consisted of the ratio of the buying price to one lottery's mathematical expectation plus the ratio of the selling price to the mathematical expectation of the other ticket. This ratio was normalized to make a value of ‘1’ equivalent to risk neutrality.

23 As a further check, individuals' scores on the index were correlated with individual's average support for the four principles. Only one weak correlation of 0.167 (significant at the 0.03 level) was found, and it was in an unexpected direction. Increasing risk propensity was positively correlated with support for Rawls's principle.

24 Although it would appear reasonable to remove knowledge of one's economic status to behind the veil of ignorance, income aspirations are a bit trickier. These latter are perhaps considered as part of a person's conception of the good by Rawls. Yet for a rational individual to choose behind the veil of ignorance, (s)he must retain knowledge of some values. Here, and elsewhere, the ambiguities of Rawls's work make for difficulty in precise simulation of the original position.

25 An ancillary test of the impact of these variables was performed by correlating them against average support for the various indices. Responses to the question asking what percentage of the student's college expenses are met by the student's parents are negatively correlated with the average levels of support for the floor constraint principle. A correlation of —0.191 with a p of 0.013 emerges. In a similar vein, a positive correlation of 0.145 (significant at the 0.045 level) is found between the support for a floor constraint and the percentage of a subject's college education which is financed via loans. No other principle is significantly correlated.

26 Again as an additional test, the income aspiration scores of the entire set of subjects were correlated with their mean support for the principles of justice. Only one correlation was found. A weak negative correlation (r = —0.151, p = 0.035, n = 145) between income aspiration and support for Maximizing the Average with a Floor Constraint was the only relationship which emerged.

27 Of course, it is not so clear whether the ideological leanings of a person should be part of the material brought behind the veil of ignorance. This is not explicitly dealt with by Rawls, and although the reader might believe that the ideological orientation of the individual could be part of the individuals' conception of the good, this is not necessarily in keeping with Rawls's notion of the ‘conception of the good’.

28 It may be that Canadian and American subjects differed in their interpretation of the connotations of the words ‘liberalism’ and ‘conservatism’. Nevertheless, the general direction of the relationship appears not to be affected by these differences in nuance.

29 Tests are available to see if expectations regarding the likely attainable floors led to the choices which were made. One could manipulate the amounts specified in the examples in the choice situations in Part I. This would allow us to see if, as the specified floor decreases, there is increased demand for floor maximizing. Frohlich, N., Oppenheimer, J. and Eavey, C., ‘Choices of Principles of Distributive Justice in Experimental Groups’, American Journal of Political ScienceGoogle Scholar, forthcoming, reports on the results of experiments with lower floors. They do not offer evidence that Rawls is correct in the fundamental reasons for the attractiveness of the mixed conceptions of justice.

30 Frohlich, et al. , ‘Beyond Economic Man’Google Scholar; Bond, Pat, ‘Game Choice Experiments and Social Interaction: Testing A Simple Model’Google Scholar, paper given at the annual meeting of Public Choice Society in Savannah, Georgia, 1983.

31 We are indebted to Oran Young for this observation.