«DiscoVeriNg HomelessNess Volume 13, Number 1 • 2011 U.S. Department of Housing and Urban Development | Office of Policy Development and Research ...»
Findings This section addresses major findings from Parts I and II of the study. Part I discusses descriptive statistics from the AHS on differences in housing quality for Blacks and Whites and then presents multivariate models predicting housing inadequacy and overcrowding. Part II discusses descriptive statistics on geographic location of the Black homeless population, the migration pattern of homeless clients, and the duration and transiency of homeless spells and then presents multivariate models predicting migration for homeless services.
AHS Descriptive Statistics Before turning to the multivariate analyses, it is necessary to have a sense of the general patterns present in our variables of interest. Exhibit 1 presents descriptive statistics by race for the full sample
and for those living in central cities. Hispanic householders and householders who identified their race as something other than Black or White were excluded from the analyses because (1) the segregation indices used in the analyses represent Black and White segregation, not Hispanic and White non-Hispanic segregation; (2) research suggests that the relationship between segregation and housing quality is different for Hispanics (Baker, 1994); and (3) this study’s primary aim was the analysis of the relationship between segregation and housing outcomes for Blacks. Regarding Blacks and Whites overall, housing inadequacy, overcrowding, and homeownership results mirror those found in previous studies. More than the majority of both Blacks and Whites live in adequate housing. Although 13 percent of Blacks live in inadequate housing, only 5.9 percent of Whites live in similar conditions. Blacks are more likely than Whites to live in overcrowded housing (4.5 percent for Blacks compared with 2.3 percent for Whites). Whites are also more likely to own their homes (69.4 percent) than are Blacks (43.1 percent). Although more than one-half (55.9 percent) of Blacks live in central-city areas, only a little more than one-fourth (25.9 percent) of Whites live in the central city. More than one-half (53.9 percent) of Black householders are female, but only 31.1 percent of White householders are female. The percentage of Blacks on welfare is more than three times the percentage of Whites on welfare (14.3 percent versus 4 percent). Blacks have higher percentages of householders whose highest educational attainment is less than college. White householders are almost two times as likely as Blacks to attain college as the highest level of education (17 percent compared with 9.9 percent, respectively). In the overall sample, households with White householders have a mean income of $46,855, and those with Black householders have a mean income of $30,123.
The mean age of White householders is 48.7 years, and the mean age of Black householders is
There is reason to expect housing quality to be worse in central-city locations than outside centralcity locations due to the concentration of poverty within inner city areas. Because Blacks are more likely than Whites to live in central-city areas, we might expect them to be more likely to live in lower quality housing. Do Whites living in similar areas also experience the same housing quality problems? Focusing on the central city section of exhibit 1, we see that both Blacks and Whites have higher percentages living in inadequate housing, but Blacks still have higher percentages in inadequate housing than Whites have (13.3 percent versus 8.2 percent). The crowding measure is very similar for Blacks and Whites in the central city, with 4.6 percent of Blacks in crowded housing and 3.6 percent of Whites in housing that is crowded. Smaller percentages of both Blacks and Whites own homes in the central city, but Whites maintain their lead over Blacks with more than one-half (53.8 percent) owning homes compared with only 34.8 percent of Blacks. Still, more than one-half of Black householders are female (56.8 percent), but only 36.5 percent of White householders are female. The percentage of Blacks and Whites receiving welfare in the central city is almost the same as in the overall sample. The percentage of Blacks living in public housing is four times the percentage of Whites living in public housing (9.1 versus 1.9 percent). Much like in the overall sample, Blacks have higher percentages than Whites who have finished less than college, but more than two times the percentage of Whites attain college as their highest level of education compared with Blacks (19.5 versus 9.3 percent). In the central city, households headed by White householders have a mean income of $43,152 and households headed by Black householders have a mean income of $27,452. The mean age of White householders is 47 years and the mean age of Black householders is 45.2 years.
44 Discovering Homelessness From Exclusion to Destitution: Race, Affordable Housing, and Homelessness
In both the overall sample and in the central city sample, Blacks experience housing inadequacy and overcrowding at higher levels than Whites do. Socioeconomically, Black householders are less likely than their White counterparts to receive college degrees and are more likely to earn less.
Blacks are more likely than Whites to live in public housing, be on welfare, and live in femaleheaded households. Although they do suggest racial differences in housing quality, these descriptive analyses do not explain the relationship between segregation and race in determining housing outcomes for Blacks. The next section of this article examines these relationships.
AHS Multivariate Analyses This section discusses the results of logistic regression models predicting housing inadequacy (exhibit 2) and overcrowding (exhibit 4). The exhibits present models for Black and White owners and renters in the 1997 AHS national sample. All models were significant at p.001 and all
$5,000–$9,999.840*** 2.316.356*** 1.428.355*** 1.426.355*** 1.426.358*** 1.430.389*** 1.476.387*** 1.473 (.004) (.005) (.005) (.005) (.005) (.005) (.005) $10,000–$14,999.733*** 2.081.507*** 1.660.506*** 1.659.506*** 1.659.508*** 1.662.534*** 1.706.530*** 1.699 (.004) (.005) (.005) (.005) (.005) (.005) (.005) $15,000–$19,999.568*** 1.765.372*** 1.451.368*** 1.445.368*** 1.445.371*** 1.449.392*** 1.480.388*** 1.474 (.004) (.005) (.005) (.005) (.005) (.005) (.005) $20,000–$29,999.477*** 1.611.240*** 1.271.239*** 1.270.239*** 1.270.242*** 1.274.277*** 1.319.273*** 1.314 (.004) (.005) (.005) (.005) (.005) (.005) (.005) $30,000–$39,999.398*** 1.489.259*** 1.296.256*** 1.292.256*** 1.292.259*** 1.296.261*** 1.298.259*** 1.296 (.004) (.005) (.005) (.005) (.005) (.005) (.005) $40,000–$49,999.271*** 1.311 –.088***.916 –.090***.914 –.090***.914 –.088***.916 –.083***.920 –.087***.917 (.004) (.005) (.005) (.005) (.005) (.005) (.005) $50,000–$74,999.106*** 1.112.048*** 1.049.046*** 1.047.046*** 1.047.047*** 1.048.050*** 1.051.047*** 1.048 (.004) (.005) (.005) (.005) (.005) (.005) (.005) $75,000–$99,999.223*** 1.250.302*** 1.353.298*** 1.347.298*** 1.347.298*** 1.347.302*** 1.353.298*** 1.347 (.004) (.005) (.005) (.005) (.005) (.005) (.005) $100,000–$124,999 –.112***.894 –.201***.818 –.205***.815 –.205***.815.205*** 1.228 –.226***.798 –.229***.795 (.004) (.006) (.006) (.006) (.006) (.006) (.006)
coefficients were significant at p.05 (most were significant at p.001). Regression coefficients are presented in the exhibits along with standard errors and odds ratios. Odds ratios, which are exponentiated regression coefficients, are discussed in the text.
Exhibit 2 shows that Black householders were 1.439 times more likely than White householders to live in inadequate housing, after controlling for other factors (model 1). Those living in central-city areas were 1.232 times more likely to live in an inadequate unit than were those living outside central-city areas. Housing inadequacy declines with increases in education, with householders who have an eighth grade education or less being 2.472 times more likely to live in an inadequate unit than householders with more than a college education. Owners are about one-half as likely as renters to live in an inadequate unit. Model 2 introduces the dummy dissimilarity measures into the model. Higher segregation rates are associated with higher levels of housing inadequacy. Units in the highest segregation quartile are 1.259 times more likely be inadequate compared with units in the lowest segregation quartile.
Black segregation interaction terms are added to model 3. At ever-increasing levels of segregation, housing inadequacy increases for the overall sample. In exhibit 3, coefficients from model 3 were used to graph the predicted probability of living in an inadequate unit for Blacks and Whites.
As we see, at low and medium levels of segregation, Blacks and Whites have similar predicted probabilities of living in inadequate units. In the highest segregation quartile, Blacks are more likely than Whites to live in inadequate units, providing some support to hypothesis 1 that high levels of segregation decrease Black housing quality. After controlling for background factors, we find that being a Black householder, living in a more segregated metropolitan area, and being a Black householder living in a highly segregated metropolitan area increase the odds that of living
50 Discovering Homelessness From Exclusion to Destitution: Race, Affordable Housing, and Homelessness in an inadequate dwelling. This finding suggests that segregation does not affect Black and White differences in housing adequacy until segregation rates are in the highest quartile.
Living in the central city also increases the odds of living in an inadequate dwelling as does renting the housing unit. This finding suggests that those renting units may have less control over the maintenance of their units, thus resulting in a greater likelihood of inadequately maintained units.
Householders receiving welfare were more likely to live in inadequate units, as were householders with less than a high school education. Compared with those living in the West, those living in the Northeast were more likely to live in inadequate units and those in the Midwest and South were less likely. Older householders had lower odds than younger householders of living in inadequate dwellings. Households with incomes of less than $5,000 were most likely to live in inadequate housing. Income had a nonlinear effect on housing inadequacy. The nonlinear effect of income is perhaps due to cost-of-living differences in different metropolitan areas not accounted for in the models. Across different metropolitan areas, the same income has different purchasing power, dependent on differences in housing costs.
In models 4 and 5, affordability measures are introduced into the models, testing the first part of hypothesis 3. As expected, high rent burdens increase the likelihood of living in an inadequate unit and a higher ratio of lowest rent units to lowest income households decreases the likelihood of living in an inadequate unit. Adding the ratio measure reduces, but does not erase, the effects of segregation on housing inadequacy for the overall sample or for Blacks in particular. This finding suggests that increasing the supply of affordable housing will mitigate but not remove the effects of segregation on the individual housing situations of poor Blacks living in the most segregated metropolitan areas.
In models 6 and 7, the effects of increasing metropolitan area homeownership on housing inadequacy are tested, the first part of hypothesis 4. Findings indicate that householders living in metropolitan areas are less likely to live in an inadequate unit if area homeownership rates are high. Adding homeownership rates to the model reverses the effects of segregation on housing inadequacy. The addition of homeownership rates reduces the effect of Black headship on housing inadequacy, but it does not erase the effect. The coefficient for Black headship decreases from.077 to.027 when homeownership rates are added. The addition of homeownership rates reduces the odds of a Black-headed housing unit being inadequate from 1.08 times to 1.027 times the odds of a White-headed housing unit being inadequate. This finding suggests that policies that promote homeownership may decrease the likelihood of living in an inadequate unit for the overall population, but this effect may not carry over to the Black population to the same extent it affects the White population. Surprisingly, controlling for homeownership rates reverses the effects of higher affordable housing supply on housing inadequacy.
Regarding overcrowding, model 1 (in exhibit 4) shows that Black householders are 1.51 times more likely than White householders to live in crowded units, even after controlling for other factors. Increasing education level greatly decreases the likelihood of living in crowded housing, with householders who have an 8th grade education or less being 15.226 times more likely to live in a crowded housing unit compared with householders with more than a college education.
Owners were about one-half times as likely as renters to live in a crowded unit. Those living in public housing were.752 times as likely as those not living in public housing to be living in