26 Mar 2013Working Paper Summaries

How Elastic Are Preferences for Redistribution? Evidence from Randomized Survey Experiments

by Ilyana Kuziemko, Michael I. Norton, Emmanuel Saez and Stefanie Stantchev

The United States has witnessed a large increase in income concentration over the past several decades. While standard theory predicts that support for redistribution should increase with income inequality, there has been little evidence of greater demand for redistribution despite historic increases in income concentration. A possible explanation is that people are unaware of the increase in inequality, and greater information would substantially move redistributive preferences. The authors examine the determinants of redistributive preferences through randomized online survey experiments with Amazon Mechanical Turk, a rapidly growing new laboratory to carry out social experiments. The results show that information changes people's beliefs about the current level of inequality and leads them to become more favorable to redistributive policies like the estate tax. Key concepts include:

Respondents update their views about income inequality when presented with information about the current distribution of income. Liberals and conservatives differ greatly in their perception of inequality, but receiving information about inequality closes roughly forty percent of this "opinion gap."
Receiving information about current levels of income inequality causes liberals and conservatives to become more supportive of redistributive policies such as the estate tax.
At the same time, emphasizing the severity of current income inequality appears to undercut respondents' willingness to trust the government to fix societal problems. The existence of the problem appears to serve as evidence of the government's limited capacity to improve outcomes more generally.
Randomized surveys are a powerful and convenient tool for studying the effects of information treatments on attitudes and behaviors.
mTurk is powerful because it can reach large samples of U.S. residents (in the thousands) at fairly low cost ($1-$2 per respondent). It is convenient because, using widely available software, online surveys are now very easy to design.
In contrast to field experiments, which are costly to set-up and replicate, online survey experiments lend themselves naturally to conducting series of experiments where results from an initial experiment lead to new experiments to cast light on potential mechanisms.