In response to Michael‘s comment, let me expand on the ethnic diversity/social welfare spending post, which I think makes an interesting point that you rarely see in mainstream media commentary.
As I read the Glaeser/Alesina study (I haven’t read their entire book), they are making a positive, not a normative argument. Essentially, they assume that racism is simply a part of human nature; that all of us are programmed — quite possibly genetically — to prefer associating with people who share our race or ethnicity. As far as I am aware, this assumption is not controversial among evolutionary biologists:
Racism is a subset of ethnocentrism, the tendency to favor genetically, socially, and culturally similar in-groups over alien out-groups. Edward O. Wilson has argued that selection favored those humans who were the quickest to recognize, fear, hate, and drive away or kill strangers, thereby securing a margin of safety for themselves and their kin. It does appear that ethnocentrism, or at least xenophobia, is in some degree biologically programmed.
[David T. Courtwright, Violent Land: Single Men and Social Disorder from the Frontier to the Inner City, Harvard Univ. Press, p. 31). Of course, this inborn tendency is often reduced by socialization, just as with other evolutionary drives (another example: about 35% of males are willing to say they would commit rape if they could be guaranteed that they’d never be identified or punished).
Surveys consistently uncover plenty of racial and ethnic prejudice across all societies, and usually disprove the contact hypothesis. That is, increased contact with people of difference races or ethnicities is just as likely to deepen your distrust of them as it is to reduce it. I can’t tell you how many Germans I’ve met who denounce anti-black racism in the United States, but then turn right around and lament the ‘inherent’ laziness and criminality of Turks or Poles. To paraphrase Czeslaw Milosz, it’s always the other country’s Indians who are the innocent victims of oppression — the ones in your own country deserved their fate.
People exist who regard race and ethnicity as irrelevant cultural constructs, but this view is widespread only among elites. (Who, in turn, sometimes have a hard time distinguishing the views of the people they meet everyday from the very different views held by majorities of their fellow citizens.) Outside of these elites, average people regard ethnic differences as very real, and very important. How do you identify who’s racially or ethnically different? Simple: they look different, they go to a different church, they have different customs, and/or they speak a different language.
You also learn how to identify ‘others’ from the your group’s collective cultural memory. Just ask any resident of a country that was once part of the Ottoman Empire what they think of ‘the Turks’, and you’ll get an earful. I’ve already tried this experiment on Bulgarians, Serbs and Greeks, and gotten very similar results. Many houses in these countries contain secret doors — which their owners will proudly show you — used for centuries to flee from marauding Turkish soldiers.
These are the not-pretty facts of human nature that explain the social welfare / civic cohesion study results. Assuming some degree of widespread ethnocentrism as a given, you would expect it to be easiest to convince people to participate in collective-welfare schemes if almost all the people who would benefit came from their ‘group.’ You would expect it to be especially hard to set up these schemes if the people who stood to gain the most from them or to gain disproportionately from them were ethnic minorities. And that’s pretty much what you see.
Thus, racism doesn’t drive the creation of social-welfare systems; it simply complicates their creation in diverse societies.