Why are levels of welfare benefits lower in some states? Research shows that racism may be one of the reasons

Why are levels of welfare benefits higher in some American states and lower in others? Some may speculate that maybe residents in the former states are altruists and more moral than others.

Academic research however shows that the differences are the results of both financial self-interest and interpersonal preferences (well, we usually call it racism).

A study done by Harvard professor Erzo Luttmer and published in the Quarterly Journal of Economics shows that:
(1) Individuals decrease their support for welfare if there are more welfare recipients in their area
(2) Individuals increase their support for welfare spending if a larger fraction of welfare recipients in their area belongs to their racial group (Question: which  word with r as the initial do we usually use to describe such type of preference, thought, and behavior?)

He derives the results from nation-wide surveys as well as voting behaviors in California’s Proposition 165 in 1992 primaries, in which governor Pete Wilson proposed both cuts in welfare generosity and changes in the state budget process.

Professor Luttmer concludes that the results also help to explain why welfare benefit levels are relatively low in racially heterogeneous states. Actually, the results also help to explain why European countries redistribute more than we do, as they are usually racially more homogenous.

Group Loyalty and the Taste for Redistribution (PDF file)
Abstract: Interpersonal preferences - preferences that depend on the characteristics of others - are typically hard to infer from observable individual behavior. As an alternative approach, this paper uses survey data to investigate interpersonal preferences. The General Social Survey contains self-reported preferences for welfare spending, which I validate with voting behavior on cuts in welfare benefits. Using this preference measure, I show that preferences for income redistribution are not only determined by financial self-interest but also by interpersonal preferences. These interpersonal preferences are characterized by a negative exposure effect - individuals decrease their support for welfare if there are more welfare recipients in their area - and racial group loyalty - individuals increase their support for welfare spending if a larger fraction of welfare recipients in their area belongs to their racial group. My results hold when areas are defined as states, metropolitan areas or census tracts and are robust to various specification checks. Direct evidence that individuals' preferences for redistribution are partly determined by the effects of redistribution on the utility or lifestyle of others in their community is valuable for the development of more accurate theoretical models and for the design of redistributive policies. The results also help to explain why welfare benefit levels are relatively low in racially heterogeneous states.