Geoscience Reference
In-Depth Information
tract household income to the city median in the same period, after controlling for
the same physical structure measures as included in Table
19.3
.
17
There is a positive relationship between the concentration of gay male house-
holds in 2000 and population growth for 2000-2009 nationally and in the Northeast.
In contrast, the share of lesbian households has no effect on population growth, and
a positive effect on income growth only in the Northeast, which is consistent with
the hypothesis that their greater likelihood of having children and their lower
incomes lead to intra-city locations different from those of gay men. This result
may also have different implications for neighborhood economic development.
Household income growth generally decreases, however, with the census tract's
share of non-family households in 2000. Non-family households include single
persons (the main component), as well as adults not related by blood or marriage
who share housing. We control for non-family households in these analyses
because, as indicated in Table
19.3
, there is a strong positive correlation between
the presence of gay male households and nonfamily households in census tracts,
and the explanations as to why a spatial concentration of gay households might lead
to greater income or population growth seemingly apply to spatial concentrations of
single people as well [see Collins (
2004
)]. In fact, Tables
19.4
and
19.5
indicate that
the census tracts with more non-family households experience less income (and
implicitly housing price) and population growth (in the Northeast and West) over
the following decade. Although non-family households include gay households,
gay households are a trivial share of non-family households (less than 3 % of
non-family households in these cities in 2000, for example). The correlations
between population or income and non-family household shares in Tables
19.4
and
19.5
reflect the relationship of population or income growth to concentrations
of single person households.
19.1.5 Conclusions, Qualifications, and Next Steps
This first attempt to analyze the locations of gay male and lesbian households in
large cities in the U.S. has several rather surprising results.
First, gay men and lesbians are less segregated within cities than African
Americans, and gay males are more segregated than lesbians. Although gay male
segregation increased slightly, on average, between 2000 and 2010 in the West,
South, and Midwest, lesbian segregation during this time decreased substantially in
all regions. Consistent with the finding of lower levels of lesbian segregation,
lesbian households are somewhat less likely than are gay male households to be
17
As in Table
19.3
, OLS estimation is performed to test for spatial dependence in the residuals and
to choose the appropriate spatial model specification. For the population growth equation a spatial
error model is estimated. For the income growth equation a spatial autoregressive model is used.
The LM statistics from the OLS estimations are displayed in Table
19.6
in Appendix.
