index of racial segregation, if the mean for the early period is significantly higher than the mean for the modern period, it can be concluded that the civil rights and fair housing legislation contributed to the reduction in the level of racial segregation and discrimination in the low-income housing program. If the mean of the two estimates of the index (Ixn) for the early period is not significantly different from the mean of the estimates for the modern period, it can be concluded that the civil rights and fair housing legislation have had little or no effect on the level of racial segregation and discrimination in subsidized low-income housing. It is not expected that the indexes would be higher for the modern period.
This test, like the one for income separation, has certain limitations. In the early period, low-rent public housing for families was the only federally subsidized housing program available to the eligible population. In the modern period, members of the eligible population could choose among an increasingly large number of housing subsidy programs each tailored to the needs of a particular clientele. Some of these housing subsidy programs like the Section 23 Leased Housing and the Section 8 Housing Assistance Program were, in part, designed explicitly to reduce the level of tenant segregation by race. The values of the second estimate of the index of racial segregation are based on data on the racial composition of households in PHAs with a total population of 50,000 or more. This means that the pattern of racial occupancy of housing projects in smaller PHAs is not taken into account in our comparative analysis.
Since a single composite index of racial segregation is calculated for diverse groups of housing subsidy programs, its complexities may make a simple comparison difficult. The results should be taken as indicative rather than definitive due to the imperfections in the data. However, since every effort has been made to render the index estimates conservative (i.e., counter to the hypotheses under consideration), some confidence can be placed in any positive outcome of the statistical tests.
In analyzing segregation in subsidized low-income housing, the question of racial occupancy patterns should be clearly distinguished from that of income dispersal. Income dispersal refers to the distribution of housing projects among income areas in the PHA. Racial occupancy refers to the distribution of racial groups of tenants across housing projects. Together, income dispersal and racial occupancy constitute the two dimensions of segregation in a subsidized low-income housing program.
For a given pattern of income dispersal of housing projects there are potentially several patterns of racial occupancy. Subsidized housing projects could be uniformly distributed among existing income areas in a PHA but remain