Does white flight still happen? Perhaps the rapid changes in racial/ethnic composition today would be better described by another term, nonetheless, rapid changes do seem to occur in many large American cities.
Here's a look at how these changes have occurred (between 2000 and 2010) in two cities, Detroit and Houston.
Detroit has been used as a textbook historic example of "white flight", having gone from overwhelming white to overwhelming non-white (mostly black) in the last century. This seems to have gone hand in hand with inner city decline, which continues today, spreading into more outlying neighbourhoods. Much of the census tracts experiencing rapid changes in racial composition are probably the result of Hispanics and especially blacks moving to these areas and out of the inner city.
Houston is rather different, while Detroit saw essentially no population change in its metropolitan area, Houston is one of the fastest growing metropolitan areas in the country. However, 93% of its population increase is the result of an increase in the population of its various visible minority groups, which now make up 59.5% of the metropolitan area's population. The 1.1 million in visible minority population added in the last decade had to be housed somewhere, creating potential for rapid changes in ethnic composition.
So lets look at which census tracts have had the share of the population that are visible minorities increase by over 20 percentages points (ex from 55% VM to 75% VM).
Starting with Detroit, with census tracts where blacks have seen the biggest increase in population share in blue, Hispanics in orange, and Asians in red.
It seems like much of the black population of Detroit is moving into neighbourhoods that are just 1-2 miles further out and only a bit newer. There are still blacks moving to areas further out THOUGH, such as Westland, Canton, Van Buren and Wixom. Other areas that experienced increases include parts of Inkster and Southfield, which has already had a black majority in 2000, and parts of Farmington Hills adjacent to Southfield.
There is a similar pattern with Hispanics, though on a much smaller scale since they make up a smaller part of the overall population, with big increases in their population share on the outer edge of the existing hispanic communities in SW Detroit and northern Pontiac.
Asians saw large increases in their population share in communities that already had relatively significant Asian populations, in Hamtramck, Troy, and rapidly growing parts of Canton.
In the case of Detroit, not all areas that experienced an increase in the share of minorities experienced a decrease in the number of whites. There were a few cases where both the white and minority population increased, but the increase in the number of minorities sufficiently outpaced the increase in the number of whites that the minority share nonetheless increased by >20%. Such areas are mapped in white below.
So how does this compare to Houston?
Which minority drives the increase follows somewhat similar patterns to Detroit.
Much of the city of Houston's Asian population are in the SW of the city, and the big increases in the share of Asians have been in the SW suburbs like Sugar Land and Cinco Ranch. Hispanics are the dominant group in SE Houston and one of the areas where their share is rapidly increasing are SE suburbs like Pasadena, South Houston. Hispanics account for 62% of Houston's population growth, so it's not surprising that they are the most rapidly growing minority group in the greatest number of neighbourhoods.
In any case, the minorities in Houston are less segregated from each other (and from whites) than in Detroit with many areas having several minorities well represented.
Here's second map for Houston (with a black outline so that the areas in question stand out more against light grey subdivisions and industrial areas).
*These areas may have seen a big decrease in the number of whites remaining, but even if they go from 5% white to 1% white, the share of whites has only decreased by 4 percentage points, much less than the 20% required to show up on the map. I don't consider this a flaw with my methodology though, since in such neighbourhoods, the racial transition is just about complete, and what I'm interested in are areas that are still changing rapidly.