Consequential egalitarianism vs. accountability principle: an experimental investigation

accountability principle: an experimental investigation Zhicheng Phil Xu and Marco A. Palma conducted a laboratory experiment to examine how people's fairness views and redistribution depend on different rooted risks. The experiment investigated two contradictory views about distributive justice, namely consequential egalitarianism and the accountability principle (AP). The results show widespread support for the view of AP when participants have the opportunity to alleviate the risks. The experiment involved three pairs of treatments and controls, differing in the potential income inequality: low-, medium-, and high-income inequality (HII). In the low-income inequality (LII) treatment, participants were informed that one of three possible outcomes would be realized with equal probability. Before the outcome was unfolded, subjects were asked to decide whether or not to buy insurance at a cost of 5 points. Outcome A did not cause financial loss, while Outcome B caused a loss that could be alleviated with the insurance. Outcome C was an 'inevitable' loss irrespective of insurance. During the second phase, participants were anonymously and randomly matched with a sequence of eight other participants. Participants with higher earnings were asked to make redistributions between counterparts. In each pair, the distributor was provided with the information about the insurance buying decision of their counterpart and the realized outcome. The computer program skipped redistribution in case of tied earnings. In the corresponding control, subjects were not provided the chance to buy insurance. Each participant was randomly assigned to one of the risky scenarios in the treatment. The experiment was computerized with z-Tree and conducted at the Economics Research Laboratory at Texas A&M University using 228 students recruited through ORSEE. A total of 78 participants were assigned into the pure-luck control, and 150 subjects participated in the option-luck treatment. Each subject participated in only one session that lasted approximately 30 minutes. The results show a sharp distinction in the redistributive transfers between the controls and treatments. Over 60% of distributors in the 'option-luck' treatments did not transfer to their counterparts at all, and about 20% of transfers were less than 20% of pre-distributed earnings. In contrast, the distributors made significantly higher transfers to counterparts in the 'pure-luck' control. About 40% of transfers were zero, while more than 40% of transfers were 20% or more of pre-distributed earnings. Table 1 provides the net payoff structure for the LII treatment. The results also show that the average transfer was 2.38 ECUs (or 15.54% out of pre-distributed earnings) in the 'pure-luck' controls, significantly higher than in the 'option-luck' treatment, 1.07 ECUs (7.77%) (p < 0.001, Mann-Whitney U-test). The sharp evidence was also found in subdivided pairs of treatments and controls. A model of fairness preferences proposed by Cappelen et al. (2007) was used to further analyze the results. The model assumes that the distributors endorse either CE or AP. The results show that the distributors who endorse AP make significantly lower transfers in the 'option-luck' treatment than in the 'pure-luck' control. This suggests that the accountability principle is widely supported among subjects, and that they are more likely to make transfers when the loss is due to pure luck rather than option luck. The experiment provides evidence for the accountability principle and consequential egalitarianism. The results show that people's fairness views and redistribution depend on different rooted risks, and that the accountability principle is widely supported among subjects. The experiment also provides insights into the fairness preferences of individuals and the factors that influence their redistributive decisions.