Homeownership rates equal the number of households that own homes divided by the number of households in the population. For that reason, differences in the propensity to form a household, or the headship rate, have the potential to explain observed changes in homeownership rates over time in addition to longstanding racial gaps in homeownership. In this regard, it should be emphasized that “headship” refers to whether an individual is identified in the Census as a household head. Thus, for example, a spouse or child of the household head would both be considered members of the household, but not the household head. We examine these questions on an age-specific basis using data from the 1970 to 2000 public use micro samples (PUMS) of the decennial census.

Summary measures indicate that age-specific homeownership rates changed little from 1990 to 2000, in contrast to the much advertised increase in aggregate homeownership rates over this period. This is consistent with evidence reported by Eggers (2004, Tables 1 and 2) when he restricts his analysis to just the relationship between household age and the propensity for homeownership using the 1990 and 2000 Censes. Looking over a longer time frame, from 1970 to 2000, age-specific homeownership rates fell by 5 percentage points for individuals from their mid-20s to mid-30s. That difference diminished thereafter, reaching zero for individuals in their mid-40s, and then rose to positive 10 percentage points among individuals in their 60s.

We find that changes in headship behavior over time contributed little to these observed patterns. For those segments of the population where changes in headship behavior did affect homeownership rates, lower headship rates reduced homeownership. This occurred because with lower headship rates some prospective households do not form, and many of these prospective households would have been owner-occupants. This pattern is most notable for individuals in their early and mid-20s for whom reductions in headship rates between 1970 and 2000 served to depress homeownership rates by 3 to 5 percentage points. That effect accounts for much of the observed decline in homeownership for this group over the 1970 to 2000 period.

Additional findings indicate that for the year 2000, black and Hispanic homeownership rates are sensitive to differences in headship behavior relative to white individuals, although primarily only for individuals in their 20s, 30s, and 40s. Among African Americans, headship rates are higher than among white individuals, and that difference serves to narrow the observed white-black gap in homeownership rates by roughly three percentage points. Among Hispanics, headship rates are lower than among white individuals, and that difference serves to widen the observed white-Hispanic gap in homeownership rates by two to three percentage points. Once again, lower headship rates are associated with lower homeownership rates. Moreover, controlling for headship behavior, white-black homeownership gaps are somewhat more severe than previously recognized, while the reverse is true for white-Hispanic gaps in homeownership.