Objective self-interest has very little effect on people’s political views.  The obvious explanation is that people vote for ideals, not objective interests.  But Gimpelson and Treisman’s evidence on systematically biased beliefs about inequality suggests another explanation: Voting is selfish but confused.  Their published regressions don’t speak to this issue, but Treisman generously agreed to run a supplemental regression to satisfy my curiosity, adding respondents’ perceived social standing to their list of variables used to predict support for redistribution.  Here’s Treisman on what found, reprinted with his permission:

[Regression results here.]

OK, here’s Table 9 from our paper, with an extra column where I’ve added the respondent’s self-placement on the 1 to 10 scale (we interpret this as being about incomes, but the question’s wording is: “In our society there are groups which tend to be towards the top and groups which tend to be towards the bottom. Below is a scale that runs from top to bottom. Where would you put yourself now on this scale?”).

Self-placement on the scale is significant. The effects of perceived inequality remain significant and similar to those in the previous column. So both a belief that one is low on the social scale and a belief that inequality is high are associated with a stronger demand for government redistribution. The country-wide shared belief on the level of inequality seems to matter more than the individual’s idiosyncratic opinions (i.e his or her divergence from the national average belief about inequality).

P.S. Note that I can’t make a causal claim based on just these data.

[Bryan Caplan again.]

As usual, it’s crucial to focus on the magnitude of the coefficients, not mere statistical significance.  Since the perceived Gini coefficient, like the actual Gini coefficient, is bounded between 0 and 1, these updated results imply that:

1. Moving an individual’s perceived status from the minimum to the maximum has the same effect on redistribution preferences as increasing the individual’s perceived Gini coefficient by .51 – a pretty large effect.

2. Moving an individual’s perceived status from the minimum to the maximum
has the same effect on redistribution preferences as increasing the average perceived Gini coefficient in his country by .13 – a modest effect. 

The story, in short, is that redistributive preferences are primarily shaped by overall political culture.  If most people in your country see a lot of inequality, you’ll probably support a lot of redistribution – even if you’re high status and don’t think there’s a lot of inequality!  Within each political culture, however, your perceptions about your own status and overall inequality are comparably important.