The changing determinants of inter‐racial home ownership disparities: New York City in the 1990s

Abstract

The 1990s witnessed significant changes in home ownership policy that were designed to reduce disparities in home ownership rates between whites and non‐whites. The study finds evidence that the policies were, indeed, effective in reducing home ownership disparities between whites and non‐whites. In this study data from the New York City Housing and Vacancy Survey (NYCHVS) are used to examine ethnic/racial home ownership rate disparities. Decompositions are performed to determine the component of home ownership rate differentials due to ethnic/racial group characteristics versus ability to translate those characteristics into home ownership. Presumably, a shift in public policy to be more ethnically and racially inclusive should reflect a change in home ownership rate disparities and, more particularly, a change in the component of those disparities attributed to differential ability to translate group characteristics into home ownership rates. Hence, this study addresses the success of recent policies to foster the opportunity of achieving the ‘American dream’ for non‐white families.

Keywords:

• home ownership
• inter‐racial disparities
• inter‐ethnic disparities
• New York City

Notes

Here, the use of the term ‘non‐white’ is not intended to have any pejorative meaning, rather, whites are used as the reference group in comparison to other ethnic and racial groups because the white group has traditionally had greater success in home ownership. Instead of stating each ethnic and racial group in relation to the white group, throughout this paper the groups will be summarised with the term ‘non‐white’.

The assumptions of unit variance in Equation 1 and a threshold level set at zero in Equation 2 are basically benign. If instead the unit variance of ν in Equation 1 were replaced by any constant variance, σ ², it would only result in the parameter estimates, β^hat , being off by some constant scalar, 1/σ, which would not affect the sign and significance of the estimates. In addition, if the threshold level of zero set in Equation 2 were changed to some other number, only the constant term, α, would be affected (Greene, 1993).

For a variety of reasons Puerto Ricans, who make up the bulk of the Hispanic population in New York City, may differ from other Hispanics in their tenure choices. To test this, the home buying trends of non‐Puerto Rican Hispanics were analysed separately from all Hispanics to determine if their experience was distinguishable from the largely Puerto Rican Hispanic population in general. The results of the decomposition exercise, available from the authors upon request, were virtually the same for non‐Puerto Rican Hispanics.

Due to space considerations native born are included but not foreign‐born black results. The foreign‐born results are available upon request.