Inequality aversion and altruistic concerns play an important role for redistributive voting that is particularly pronounced for above-median income earners

Other-regarding Preferences and Redistributive Politics. Thomas Epper, Ernst Fehr and Julien Senn (2020), Working paper series / Department of Economics Working Paper No. 339. https://www.econ.uzh.ch/static/release/workingpapers.php?id=1018

Abstract: Increasing inequality and associated egalitarian sentiments have again put redistribution on the political agenda. Support for redistribution may also be affected by altruistic and egalitarian preferences, but knowledge about the distribution of these preferences in the broader population and how they relate to political support for redistributive policies is still scarce. In this paper, we take advantage of Swiss direct democracy, where people voted several times in national plebiscites on strongly redistributive policies, to study the link between other-regarding preferences and support for redistribution in a broad sample of the Swiss population. Based on a recently developed non-parametric clustering procedure, we identify three disjunct groups of individuals with fundamentally different other-regarding preferences: (i) a large share of inequality averse people, (ii) a somewhat smaller yet still large share of people with an altruistic concern for social welfare and the worse off, and (iii) a considerable minority of primarily selfish individuals. Controlling for a large number of determinants of support for redistribution, we document that inequality aversion and altruistic concerns play an important role for redistributive voting that is particularly pronounced for above-median income earners. However, the role of these motives differs depending on the nature of redistributive proposals. Inequality aversion has large and robust effects in plebiscites that demand income reductions for the rich, while altruistic concerns play no significant role in these plebiscites.

Keywords: Social preferences, altruism, inequality aversion, preference heterogeneity, demand for redistribution
JEL Classification: D31, D72, H23, H24


inequality averse individuals (comprising 50% of our population), individuals with altruistic concerns about social welfare and those worse off( 35%) and predominantly selfish individuals ( 15%)


5 Summary and conclusions

Rising inequality in advanced capitalist countries has again put the issue of redistribution on
the political agenda. In this paper, we examined the role of other-regarding preferences for
individuals’ support for redistribution – a question that has so far received relatively scarce
attention in the political economy literature. To answer this question, we took advantage of
Swiss direct democracy where 4 radically redistributive proposals were put to vote in national
plebiscites during the last 10 years. This enabled us to measure people’s support for policy
proposals that were actually put to vote instead of using more general hypothetical questions
related to demand for redistribution.
Previous research suggests that other-regarding preference may have multiple facets –
i.e., individuals may not simply differ in their degree of “other-regardingness”, but that there
may be qualitatively distinct, and in some sense fundamentally incompatible, types of otherregarding
preferences. In our context, this incompatibility concerns, for example, the extent
to which other-regarding individuals are willing to sacrifice their own payoff for the sake of
achieving equality by reducing richer people’s income. Therefore, the first task is to identify
which fundamentally distinct social preference types exist in the broader population and to
assess their quantitative importance.
For this purpose, we designed an experiment that enables us to identify the existence
of distinct social preference types and their quantitative importance in a broad sample of
the Swiss population. Applying a novel Bayesian non-parametric clustering method to the
data of this experiment, we uncover three fundamentally distinct social preference types with
a clear behavioral interpretation: inequality averse individuals (comprising 50% of our
population), individuals with altruistic concerns about social welfare and those worse off
( 35%) and predominantly selfish individuals ( 15%). Interestingly, the individual-level
behavioral variation within types is generally relatively low but within the social welfare type
there are two meaningful subgroups – a strong type that puts more weight on helping those
who are worse off, and a moderate type that puts more weight on joint payoffs.
We link individuals’ type of social preference with their political support for redistribution
and show that both types of other-regarding preference are associated with a significantly
higher support for redistribution compared to the predominantly selfish type. This
association is robust to controlling for additional covariates which includes a large battery of
socio-demographic variables and other important determinants of demand for redistribution
that were previously discussed in the literature. Even after controlling for individuals’ political
identity, other-regarding preferences remain strongly associated with political support
for redistribution. In addition, we also show that social preferences are particularly strong
predictors of support for redistribution among individuals with an income above the median.
Inequality averse above-median income earners are 20 percentage points more likely
to support redistribution than predominantly selfish individuals. Similarly, above-median
income earners with a social welfare concern are 13 percentage points more likely to support
redistribution compared to predominantly selfish individuals. In contrast, for below-median
income earners the discernible effect of social preferences is strongly diluted. Finally, the
identification of two quantitatively important social preference types enables us to examine
their potentially differential role for different types of redistributive policies. It turns out that
inequality averse individuals are substantially more likely to support policies that “reduce the
income of the rich” than those with an altruistic concern for social welfare, while the latter
appear to be (slightly) more supportive of policies that “help the worse off”.
Altogether, these results suggest that one can gain interesting new insights into the political
economy of support for redistribution by taking other-regarding preferences – and the
variety thereof – into account. We therefore believe that the future research in this domain
would benefit from routinely measuring other-regarding preferences like inequality aversion
and concerns for social welfare. To make this possible, we provide a simplified version of our
experimental tool which allows the identification of the different social preference types with
only 5 different budget lines. We hope that this simplified tool will facilitate the application
of the methods used in this paper to examine the distribution of other-regarding preferences
in many more contexts including other cultures, countries and other types of redistributive
policies.