(February 2009) On Feb. 25, 2009, John Iceland, professor of sociology and demography at Pennsylvania State University, led a PRB Policy Seminar on racial and ethnic segregation in U.S. cities. According to Iceland, we are witnessing a pattern of "spatial assimilation" over time and across generations—a reduction in differences in the residential patterns across groups. He notes, however, that the extent and pace of spatial assimilation among immigrants is nevertheless affected by their race and ethnicity.

His recent book,Where We Live Now: Immigration and Race in the United States, looks at residential segregation and examines trends among various groups using a variety of measures.

After his presentation, Iceland answered a few questions about his research:

How has immigration changed the racial and ethnic composition of U.S. cities over the past 20 to 30 years?

This particular paper ["Immigrant Residential Segregation in U.S. Metropolitan Areas, 1990–2000"] focuses mainly on the 1990 to 2000 period, but certainly the longer trend has been [that of] increasing immigration. Particularly since the change of laws in 1965, we've seen a diversification of the immigration stream [to include] a wide array of countries in Asia and Africa and of course a number of Latin American countries. That trend only increased in the 90s, and immigration remains at historically high levels in the 2000s as well.

Your research focuses on segregation between foreign-born immigrant communities and more established immigrant and native-born communities.

Yeah, I'm looking by race and ethnicity at the segregation of different groups...from non-Hispanic whites. I'm asking, are native-born of a particular group, say Asians, less segregated than the foreign-born Asians? I use this kind of analysis as a window on the immigrant incorporation process.

Can you elaborate on the immigrant incorporation process?

Sure. One might think that new immigrants often settle into ethnic enclaves. But perhaps over time and across generations, they would move out of ethnic enclaves into other kinds of neighborhoods...[I'm] looking at residential incorporation in that sense...to what extent are they becoming less segregated from whites? In particular, [I'm] looking...by year of entry among immigrants but also comparing, for example, foreign-born Asians from native-born Asians.

Your paper finds that levels of segregation are much higher for foreign-born black immigrants than foreign-born Asian, Hispanic, and white immigrants. What accounts for this high level of segregation?

For one, in general, new immigrants and black immigrants tend to be relatively recent immigrants, and tend to settle in ethnic enclaves. So, other things being equal, more recent immigrants would be fairly highly segregated. I think if you then overlay this with race, as we know, the U.S. has long been characterized by a black/white divide and...certainly we still see high levels of black/white segregation. So, race plays a role for black immigrants as well. How does poverty play into it?

Both this paper and other work have indicated that poorer immigrants tend to be more segregated. I think the extent to which immigrant groups are comprised of lower socioeconomic status members...would be important...more so for some groups than others because the average socioeconomic profile is different. Black immigrants are not particularly of low socioeconomic status. Thus, because they tend to have higher incomes than native-born blacks, their socioeconomic status really cannot explain their high level of segregation. Socioeconomic status would be more important for say, Latinos, where you see a lot of bifurcation and a large segment of lower-skilled, lower-income immigrants.

So recent immigrants who are more educated tend to be less segregated than poorer immigrants?

That's right.

Can you explain the distinction between the dissimilarity index and the isolation index?

The word "segregation" can mean many things to different people. People who have studied segregation have developed different dimensions of segregation which tap into different kinds of residential patterns. Dissimilarity index is a measure that looks at how evenly groups are spread across metropolitan areas, however many members of a particular group there may be. I gave the example in the talk that if a metropolitan area is 20 percent black, for there to be no segregation according to the dissimilarity index, every neighborhood would have to be 20 percent black.

The isolation index, in contrast, measures to what extent ethnic group members live with co-ethnics. Other things being equal, when there are a lot of co-ethnics, let's say Mexicans, Hispanics, or however you want to define the group of interest, if there are a lot of Mexicans living in the metropolitan area, isolation levels, other things being equal, will tend to be higher than in metropolitan areas where, let's say an extreme case, where there's only one Mexican living in a metropolitan area. In terms of calculating a segregation index, it would be extremely low because there's mainly contact with majority group members. He or she would be mostly surrounded by majority group households. So isolation just indicates the average proportion of co-ethnics living in a neighborhood where the typical person [is also of that] particular ethnic group.

How is 'neighborhood' defined?

We use census tracts as approximating neighborhoods, which is the most common choice. They are a bit on the large side – 4,000 people or so. If you measure neighborhoods on smaller levels, places tend to look more homogeneous. But metropolitan areas that have high levels of census tract-based segregation will have high levels of block-level segregation. Those are correlated.

Can you explain regional differences and the examples of specific metropolitan areas?

Regionally, segregation levels tend to be highest in the Northeast and Midwest cities, often what are termed Rust Belt cities. These are places...such as Chicago, Milwaukee, Buffalo, New York City even...we also see relatively high levels of segregation for immigrants. Segregation levels do tend to be lower in western and southern metropolitan areas...it's not always uniformly [different across regions]. For example, Los Angeles has pretty high levels of Hispanic/white segregation.

What are the implications of your research findings and what are the policy implications?

I tend to think that in the short-run, because we have high levels of immigration, we'll see perhaps even increases in Hispanic and Asian segregation from whites. Immigrants often settle in ethnic enclaves. But my research suggests that over the longer run, we do see that immigrants become residentially incorporated into American cities...In general we see higher socioeconomic status groups tend to be less segregated than lower ones...Of course when setting policy, there are always many different considerations including humanitarian ones. The U.S. has been a country that has accepted a wide range of immigrants...Bring us your...

Your tired, your needy, your huddled masses…

Right exactly. Though it does look like the incorporation process is [easier] for higher socioeconomic status immigrants. We do see even among immigrants who come with lower levels of socioeconomic status, educational levels are higher for the second generations and beyond...We often see that there is some rise in socioeconomic status across generations, particularly for Asians. For Hispanic immigrants, this is still unfolding since immigration from non-European countries is still relatively recent.

This is largely what has happened throughout American history, looking for instance at the experience of Italian or Irish immigrants 100 years ago.

Yes. Italians came in as a relatively highly segregated group with low levels of human capital, low skill levels. It took more than one or two generations, it's really three, four, and more until they became just another white ethnic group on par with others.

Will you being continuing this research with the 2010 Census?

Just like the 1990s was a period of high immigration, immigration has continued in the 2000s, so I think we need to continue to learn about how immigrants are faring in the U.S. The 2010 Census gives us an opportunity to look at how immigrants are doing...at the neighborhood level. Between censuses, you can get a national picture of how people are doing but you don't get that geographic detail that you get with a census.


Eric Zuehlke is an editor at the Population Reference Bureau.