William A. Testa and William Sander

April 19, 2009 Abstract

     Selected  cities have become increasingly attractive to high human capital households.  This is partly a result of the location of human capital intensive industries that locate in cities with more skilled workers as well as high human capital consumers demanding the amenities of cities.  Further, it has been shown that concentrations of human capital in cities can have positive effects on the skills of workers and their earning ability thus increasing the incentive to  live in cities.     One of the consequences of the growth of human capital in at least some cities is rising economic inequality within metropolitan areas.  One of the consequences of rising inequality in cities is an increase in spatial polarization as well.  This is the case in Chicago.  Locations around the core in Chicago that had some of the highest concentrations of poverty as recently as the 1980s now have some of the highest concentrations of high human capital households.       This paper further highlights the effects of changes in education and demographic background on the location of households in the Chicago metropolitan area.  Particular attention is given to changes in (Lake Michigan) lakefront locations by households in the city of Chicago relative to household locations in the interior of Chicago and  suburban areas of Chicago (Illinois part).     We show that high levels of educational attainment are associated with living on Chicago’s lakefront relative to the rest of the city of Chicago and suburban areas.  This is not only the case for non-Hispanic whites, it is also holds for African-Americans and Hispanics.  The results suggest that highly educated blacks and Hispanics are not necessarily averse to living in the city of Chicago.

     Data from the five percent PUMS (public use microdata sample)  sample from the 2000  Census of Population are used to estimate household location in the Chicago metropolitan area (Illinois part).  Estimates were also undertaken with the five percent pums sample from the 1990 Census of Population and the smaller 2006 American Community Survey.  The results in all three samples were very similar.  Thus, we only focus on one sample. 

     Multinomial logit estimates are undertaken for all respondents twenty-five and older and for workers twenty-five and older.  Four locations are estimated in the model based upon the PUMAs (public use microdata areas) for the Chicago area.  PUMAs are sample areas with a population of at least 100,000.  The four areas include PUMAs on the city of Chicago Lake Michigan lakefront from the downtown northward (called “LakefrontN”),

PUMAs on the city of Chicago lakefront south of the downtown (called “LakefrontS”),  and the rest of  the city of Chicago (relative to suburbs).