Despite the major cultural and demographic changes in the United States over the last half century, homeownership remains a firm part of the American Dream. According to a spring 1994 Fannie Mae survey, 86 percent of Americans believe one is better off owning a home and 74 percent believe that one should buy a home as soon as one can afford it regardless of marital status or whether or not one has children. In this edition, U.S. Housing Market Conditions examines trends in homeownership from 1890 to 2000.
Past
The decennial census of 1890 was the first to ask basic housing questions and, in particular, whether one owned or rented. The census data since 1890 show three distinct eras of homeownership in America.
In the 1890-1940 period, the homeownership rate fluctuated in the 43- to 48-percent range. From 1890 to 1920, the homeownership rate fell as immigration and urbanization offset the rise in income. Income growth increased the homeownership rate during the 1920s, but the Depression more than wiped out this gain so that the rate had fallen to a low of 43.6 percent by 1940.
During the 1940-1960 period, the homeownership rate rose by over 18 percentage points, from 43.6 to 61.9 percent. This remarkable transformation was facilitated by higher incomes, a large percentage of households being in prime homebuying age groups, the FHA-led revolution in mortgage financing, the GI bill of rights, improved interurban transportation, and development of large-scale housing subdivisions with affordable houses. While all of these factors played an important role in making the United States a Nation of homeowners, it is important to note that a Department of Labor study (cited in the Housing and Home Finance Agency's Housing Statistics Handbook of 1948) reported a 53.2-percent homeownership rate for 1945. If this survey was correct, then approximately half of this change took place prior to many of these factors becoming fully effective and during a time when wartime needs virtually halted residential construction. Higher wartime incomes, the absence of many competing consumer goods, and shortages of rental housing may explain this wave of homebuying.
Since 1960 the homeownership rate has remained in the 61- to 65-percent range. After slow growth from 1960 to 1980, the rate fell to 63.9 percent in 1990. Part of the decline between the 1980 and 1990 censuses can be explained by the undercount adjustment, a first-time ever adjustment by the Census Bureau. Without the undercount adjustment, the 1990 census would show a 64.2-percent homeownership rate. An important factor in explaining the trend over this period was the virtual absence of growth in real family income. Between 1980 and 1992, median family income grew only 2.7 percent in real terms.
The Current Population Survey (CPS) shows that the homeownership rate declined slowly but steadily during the 1980s before stabilizing in the early 1990s (see Table 25 in the Historical Data section). The CPS reports a 63.8-percent home-ownership rate for the second quarter of 1994.1
Present
Several facets of current trends in homeownership are noteworthy. First, the homeownership rate declined among many segments of the population. For example, Table 25 shows that the 1980 to 1993 national pattern is roughly duplicated by the pattern in the homeownership rates for all age groups except householders 65 and older.
Second, blacks and Hispanics continue to have much lower homeownership rates than whites. The 1991 American Housing Survey (AHS) reports homeownership rates of 67.9 percent for whites, 42.8 percent for blacks, and 38.8 percent for Hispanics. Large racial differences are not new; the first AHS in 1973 found homeownership rates of 67.1 percent for whites, 43.4 percent for blacks, and 43.2 percent for Hispanics. While the white and black rates remained roughly stable over this period, the Hispanic rate dropped.
Third, homeownership rates for married couples remained relatively stable between 1982 and 1993, but both single males with children and single females with children had substantial drops in their homeownership rates. At the same time, nonfamily households (single males, single females, and other two-person households) experienced rising homeownership rates.
Future
HUD's Office of Policy Development and Research (PD&R) used the 1991 AHS and recent household projections by George S. Masnick and Nancy McArdle of the Joint Center for Housing Studies to estimate what homeownership rates might be in the year 2000. Masnick and McArdle estimated the number of households in 2000 for each of 35 population categories based on new Census Bureau projections of population growth that take into account higher immigration rates than those assumed in previous census projections. The 35 categories are formed by 7 age classes and 5 family-type categories. Using data from the AHS, HUD divided the Masnick-McArdle household estimates for each of the 35 categories into white and minority components, which were further subdivided into 5 income classes. The 1991 AHS homeownership rates for all the subgroups within each category were applied to the appropriate household estimates to calculate the number of homeowners.
Using this method PD&R calculated a homeownership rate in 2000 of 65.0 percent and a total of 68,195,000 homeowners out of 104,977,000 households. Based on the CPS estimate of 62,684,000 homeowners in the second quarter of 1994, this would be an additional 5,511,000 homeowners by 2000.
This approach can be expanded to examine barriers to increased homeownership and explore the limits to possible progress. Unless the United States experiences unprecedented cultural or economic changes, it is unlikely that there will be any significant shifts in the population among the 35 age and family-type categories estimated by Masnick and McArdle. Therefore, any significant improvement in the national homeownership rate would have to derive from increases in the homeownership rates for the individual categories.
With this in mind, PD&R analyzed each of the 35 age and family-type categories that were divided further into 350 subgroups based on race (white versus minority) and income (5 income classes). Within almost every category defined by age and family type, homeownership rates rise with income and are higher, at almost every income level, for whites than for minorities. The following table shows the typical pattern, using married couples with children with a head between 35 and 44 years old as an example.
This analysis highlights the importance of income/affordability and race/ethnicity in determining homeownership rates. Under current economic conditions, whites in the $80,000+ income class probably have the highest homeownership rate consistent with the characteristics of a particular age and family-type category. Based on this premise, the aggregate homeownership rate for a particular age and family-type category can be increased only by reducing the income or minority differentials in homeownership rates within that category.
The following table simulates what the national homeownership rate would be in 2000 if the income/affordability and racial/ethnic barriers could be reduced by varying degrees. For example, in the context of the married couples, 35-44 age category, a 50-percent reduction in the racial and the income differentials would have the following effects: the white $20,000-39,000 rate would rise to 87.5 percent, the minority $80,000+ rate would increase to 89.5 percent, and the minority $20,000-39,000 rate would get both an income and a racial adjustment to 78.75 percent. In the table, moving across the rows from left to right indicates the effect on the national homeownership rate of reductions in the racial differentials in 10-percent intervals; moving down columns shows how much the national homeownership rate would increase with a narrowing of differentials across income groups.
The four corners of this table have specific meanings.
- The upper left corner is PD&R's 65.0 percent projection assuming no improvement in the racial and income differentials.
- The lower right corner (85.2 percent) is the rate that might be achieved by complete elimination of income and race differences within each age and family-type subgroup. Further substantial increases are possible only through other types of changes, such as a major increase in income, major new tax incentives for ownership, or other major reductions in the price of owner-occupied housing relative to rental housing.
- The upper right corner is the rate achieved by eliminating racial differences but no income differences. This effect is small because minorities will compose only 22 percent of the households in 2000.
- The lower left corner is the rate achieved by reducing all the income-related differences but none of the racial differences. Progress along this dimension produces much larger increases in the homeownership rate than equal progress along the race dimension because progress along the income dimension affects a much larger proportion (91 percent) of the population.
While the preceding table shows a wide range of possibilities, the realistic set of possibilities is much narrower. Large reductions in the race/ethnicity and income/affordability differences would be very difficult to achieve for many reasons.
HUD has already embarked on several approaches to reducing racial differences in homebuying through more aggressive enforcement of fair housing laws and fair lending requirements, and assertiveness in our oversight responsibility for government sponsored enterprises (GSE). Nevertheless, the racial differentials incorporate both current discrimination and the effects of past discrimination. While discrimination in housing and mortgage markets can be reduced, past experience shows that progress is slow. Moreover, racial differentials would persist even if there were no current discrimination because minorities have lower levels of wealth, as a result of past discrimination in education and employment.
Progress in reducing differences in homeownership rates related to income is constrained both by limits on the effectiveness of policies to promote homeownership among low-income families and by the weaker financial incentives provided to a large segment of the population. Technological improvements can lower the cost of housing production and local government initiatives can lower the cost of land by eliminating unnecessary land use controls. But there are practical limits to these gains and commuting costs constrain development of low-cost land at the urban fringe.
Housing counseling and efforts to simplify housing transactions can make homebuying easier for the uninitiated, and new technologies and new ways of doing business can make closing costs less expensive, but it is unclear how much progress can be made in this way. Subsidies and underwriting changes to make purchasing and financing easier face serious budget and cost constraints.
Homeownership rates below 50 percent occur among the youngest households (those with a head younger than 25) at all income ranges and among the low-income, nonelderly households (those with incomes below $20,000 and younger than 55 and those with incomes below $40,000 and younger than 35). These groups have much weaker incentives to purchase a home. The young are typically very mobile and renting is often a cheaper alternative for those who expect to change residence in the near future. Because of lower marginal tax rates and the large standard deduction, the tax incentives to become homeowners are much weaker for lower income households. These groups with weak incentives will compose 25 percent of the population in the year 2000.
PD&R's projections indicate that a modest increase in the homeownership rate to 65 percent will occur as the result of changing demographic conditions. Further increases are possible if racial/ethnic and income/affordability barriers can be reduced. However, large increases beyond 65 percent are unlikely without substantial institutional changes.
