The Wealth Distribution and Redistributive Preferences: Evidence from a Randomized Survey Experiment

We analyze a large-scale randomized experiment on redistributive preferences within the Austrian part of one of the most comprehensive wealth surveys - the Eurosystem Household Finance and Consumption Survey. Austria displays a nearly perfect laboratory for such an experiment as it has very low levels of wealth taxation and no inheritance tax but at the same time a rather high level of wealth inequality. We estimate the causal effect of information of one's own rank in the wealth distribution on preference for wealth taxation. Previous literature has mostly focused on the income distribution instead of wealth. We find the average treatment effect to be very small and insignifcant. For the group however, who overestimates their own position in the wealth distribution information on their true rank has a strong positive effect, while for the group underestimating their position originally the effect turns out to be negative. Both combined show up as the null effect overall. As theory suggests, information thus has a different effect depending on prior beliefs.