Experimental Design
We manipulate respondents' perception about their relative wealth. Specifically, we ask respondents about their net wealth and randomly vary the response categories of the question.3 Half of the respondents see response categories with wide intervals, e.g., the lowest category ranges up to 275,000 euro (treatment condition). Accordingly, the overwhelming majority of respondents (about 80 percent) should place themselves into this lowest category, and thus perceive their relative wealth rank as low. The other half of respondents receive response categories with small intervals where, for example, the lowest category was less than 2,500 euro (control condition). This implies, in contrast to the treatment condition, that the majority of respondents should locate themselves in the middle and top categories and thus perceive their wealth rank as higher. Subsequently, we measure respondents' risk tolerance through an incentivized lottery task.