Saturday, March 22, 2008

The Math of Ethnoreligious Conflict

Barack Obama, contrary to diversity-industry pieties, has not done as well among Hispanic voters as perhaps his campaign had hoped. The reasons why are suggestive for the mathematics of ethnoreligious conflict.

When the senator took approximately half the Hispanic vote in Virginia, commentators speculated that he had made fatal inroads into Senator Clinton's coalition, which relied heavily on white women and Hispanics. (Why it should be built on these genetic lines rather than, say, class or ideology is another question entirely.) But in fact this success was far from universal . While he did well in Virginia, which has relatively few Hispanics, in places with more, such as New Mexico, California and Texas, he has done worse. Why such a different pattern? In part because the larger proportion of Hispanics in those states forces them, under certain circumstances, to vote as Hispanics rather than as other things, a trait they share with other minority groups.

People carelessly refer to the percentage of "minorities" in the population, but it turns out the the distribution of minorities among various groups is an important predictor of ethnoreligious tension. Economists have a figure called the Herfindahl index, which is used to characterize the level of concentration in a market. The formula is Σpi2, where pi is the market share of firm i. (The pi thus all sum to one.) A value close to one means a high degree of concentration - one firm controls most of the market. A value close to zero means no firm controls much of the market, so there is a great deal of competition.

The same index can be used to measure ethnic concentration, with the pi now referring to the share of a particular demographic group in the population. A value close to one means one group dominates the population (in Korea, for example), and a value close to zero means no group has anything close to majority control. So which distribution generates the most conflict? The answer, some research suggests, is not a value close to one or zero but a value in the middle. When there is one overwhelmingly dominant group, minorities are too small to present a threat to the majority, which thus is happy to extend them equal opportunity. The minority, similarly, knows it has no chance of meaningful political participation and thus encourages assimilation to the majority values. China, for example, has 55 minority groups, but collectively they only amount to about six percent of the population. The Han majority thus looks on them with tolerance. The two exceptions, Uighurs in Xinjiang and Tibetans, actually prove the rule because they are so highly concentrated in one place, which requires the Chinese government to move in Han people to cement control, which engenders resistance among the aboriginals. For a population with many small groups, similarly, political power is unattainable. In principal there can be coalitions, but they are often impractical.

But if, say, one group is two-thirds and one is one-third, then trouble arises. Now, the minority group is big enough to exercise meaningful political power, hence to demand rents from the government. And the majority group is thus threatened. The groundwork for conflict is laid. If the minority has historically been in control (as with Sunni Arabs in Iraq), the problems are all the greater. And so in many states Hispanics now pose a threat to the political power traditionally exercised by blacks, but in isolation neither is strong enough to pose a multiethnic threat to whites. Hence, Sen. Obama's Hispanic troubles.

This approach also explains the attempts to artificially stitch together larger groups - to add together two or more small pi into a bigger pi. Hence, the insistence by some on referring not just to Japanese or Chinese but "Asian" identity, and the insistence on treating all Hispanics or Latinos as the same, a pattern I have noted before. Even though a Cuban-American in New Jersey may feel little sense of common culture with a Chicano farm worker in Watsonville, representatives of nascent Hispanic pressure groups are compelled to try to stitch all the groups together, no matter how ill the fit. They all have Spanish last names after all (and sometimes, as in the case of Bill Richardson, not even that), and that will do in a pinch. Whites have done this too, with Jews and Italians, among others, now included in a category, "white," that in common American usage once excluded them.

We can also get some sense of Europe's troubles with its immigrants. In France, Mahgreb and West African immigrants are disproportionate, as are Pakistanis in Britain. Having a more diverse mix, with no dominant group, leaves no group feeling like it is just barely on the outside looking in when it comes to politics. In the US, the vast array of immigrant diversity is thus a blessing, though the primary possibility for conflict is presented by tensions among blacks, whites and Hispanics (who are mostly Mexicans and Central Americans) in the Southwest. It is thus possible that intergroup conflict there will get worse before it gets better.

Fortunately, these things are not cast in stone. Casual empiricism suggests that the large Hispanic population in California is enmeshed in more conflict with other groups than is true in Texas. The former is where the modern ethnic studies-industrial complex is most entrenched in the universities, and where a proposition that banned illegal immigrants from receiving public services passed in the last decade. (Courts later overturned it.) This may have something to do with the greater rate of economic freedom in Texas, where taxes tend to be lower and limitations on business creation and expansion fewer. As I argue so frequently, giving people incentives to work across group lines gives them less reason to fight across them. But in countries like France, where economic stagnation and ethnic concentration among immigrants reinforce one another, the future does not seem so bright.



Post a Comment

<< Home