For many years the federal government has promoted homeownership as an important goal for low-income families. A primary motivation of this policy goal is the concept that owner-occupied housing can be an important means of wealth accumulation, particularly for those lower-income and minority families that are able to purchase homes. However, very little has been done in the housing literature to determine the importance of housing and non-housing sources of wealth accumulation. This determination has been difficult to address for three reasons. First, detailed wealth information on families is seldom available on a consistent basis. Second, such information on wealth is even less likely to be available over time so that changes in wealth can be observed. Third, the process of housing wealth accumulation is dynamic. Housing wealth accumulation depends critically on how soon a family that is renting becomes a homeowner, whether or not the family graduates to more highly valued owned units over time, or becomes a renter again and never regains homeownership.

The issues above are addressed through an analysis based upon a unique panel data set, the Panel Study of Income Dynamics (PSID) collected by the Survey Research Center at the University of Michigan. The PSID is unique because the location of households in the sample can be identified at the Census tract level. Thus, household information can be merged with Census data for the purpose of the analysis. In addition, the PSID has detailed information for each household within the sample on housing tenure (own versus rent), housing expenditure levels, and household characters, including income and a detailed breakdown of the net wealth position of the family. Because of its panel nature, families can be followed over time and changes in these factors observed.

The dynamics of housing choice and housing expenditure are modeled to predict potential housing wealth accumulation for households across income and racial groups. Specifically, a probability model is developed from which the cumulative likelihood of homeownership is calculated over time for all households in the sample. It is important to note that explicitly accounted for in this approach is the likelihood that, having become owners, households may subsequently transition back to rental tenure and/or may move to other owned units over time. A housing expenditure equation is estimated to predict, at a given point in time, expected household expenditure. The higher the likelihood of choosing ownership and the greater the expenditure predicted, the larger the household’s base of potential housing wealth. One additional element that is required to project the amount of housing wealth accumulation a given household might anticipate is the expected rate of appreciation in house value. To estimate the expected appreciation in house value, changes in the median value of owned homes in the Census tracts actually occupied by the PSID households are used (rather than regional averages). The predictions of housing wealth accumulation are compared to actual non-housing wealth accumulation during the sample period and implications drawn as to the relative importance of these two mechanisms for family wealth accumulation. The combination of the dynamic probability modeling in conjunction with the detailed geographic and wealth information represents a substantial extension of the existing literature in this area. There are a number of interesting findings in the analysis. First, it is important to note that owners often transition back to renting and, particularly among low-income minority families, do not regain owner-occupied housing. Specifically, for those low-income minority residents who transition out of owning only 37% return to owned status. For high-income white households this percentage is approximately 58%. This is a critical issue in that housing wealth accumulation is impacted by both whether or not a family returns to homeownership and how quickly this process occurs. Second, there are significant differences in the movement to a new house (typically of higher value) with associated impacts on housing wealth accumulation. For low-income minority households only 22% actually transition to a second home and of those about 14% move to a third owned home during the observation period. For high-income white families the percentages are higher, namely 33% and 28% respectively. Third, and not surprisingly, housing expenditures, the basis of housing wealth accumulation (through appreciation), also vary greatly across household types. For families differentiated by income, median house value is approximately $80,000 for high-income white households and $32,000 for low-income minorities. Finally, the impact of simple appreciation on housing wealth is relatively similar in this sample. Interestingly, appreciation between the 1990 Census and 2000 Census in owner-occupied units for tracts in which PSID families actually lived (rather than regional averages) suggests similar median appreciation rates, ranging from a high of 4.6% to a low of 4.3%.

The factors discussed above predict housing wealth accumulation estimates for families in the sample that are strikingly different across income and racial groups. For high-income white families average annual housing wealth accumulation due to appreciation (ignoring the equity down-payment and forced savings through amortization) is $4,460 dollars for high-income white households and $1,712 for low-income minority families. It is critical to recognize that the numbers for annual housing wealth accumulation compare very favorably to the actual accumulation of non-housing wealth by families over the same period. For high-income white households the average median level of non-housing wealth accumulation is $2,650, while for low-income minority household’s it is, quite simply, $0.

In terms of lower income households, non-housing wealth accumulation is at best minor and, for minority families, often negative. Thus, over the nine year period of the study, owned housing is an important means of wealth accumulation. Indeed, the results may be broadly interpreted for lower income households as implying that housing wealth is total wealth.

These results tend to support public policies aimed at both increasing homeownership opportunities in general and those policies that focus on homeownership for lower income households. Even though homeownership is not a guarantee of successful wealth accumulation, in that a small percentage where all family types lose money on their homes is observed, in general household wealth appears to be positively impacted by homeownership. This conclusion is reinforced with comparisons to accumulation of non-housing wealth. Wealth accumulation for low-income and minority households, although low, experiences a major increase through home ownership. In this regard, current initiatives to increase low-income homeownership seem both desirable and valid. Moreover, this work suggests that policies designed to ensure that once households achieve homeownership, they remain homeowners (rather than reverting to rental tenure), and policies that enable families to transition to higher valued owned units over time will increase substantially their potential housing wealth accumulation. These conclusions about the value of owned housing are reinforced when the positive social impacts of homeownership on families, particularly children, are also considered.