Patterns of Homeownership

On November 5, President Clinton announced a new initiative,involving the Federal Government, private corporations, realtors,homebuilders, lenders, insurers, and nonprofits, to raise the homeownership rate to an all-time high. The President directed HUD Secretary Cisneros, in cooperation with these groups, to develop within 6 months "a detailed strategy that recommits America to homeownership."

In its August 1994 issue, U.S. Housing Market Conditions examined trends in the national average homeownership rate and predicted a 65-percent rate in the year 2000. This prediction assumed that homeownership patterns observed in 1991 would continue as the population aged and household composition changed. (The current rate, third quarter of 1994, is 64.1 percent, the same rate as in 1991.) This issue discusses patterns within the overall rate.

The suitability of any given overall rate depends in large measure on how that rate is distributed among various segments of the population. The data presented in this discussion indicate wide variations around the national average.

Demographic Patterns

Homeownership rates vary substantially by race, income, age, family status, and other demographic characteristics. The 1991 American Housing Survey (AHS) shows how much variation there was around the 1991 national average of 64.1 percent.

• Whites had much higher homeownership rates than either blacks or Hispanicsþ67.9 percent versus 42.8 percent and 38.8 percent in 1991.
• Forty-eight percent of households with incomes less than $20,000 in 1991 owned homes. Those households with incomes between $20,000 and $39,999 had a homeownership rate, 62 percent, near the overall rate. Every higher income group had a homeownership rate substantially greater than the overall rate, with 90 percent of those with incomes of $80,000 or more owning homes.
• Table 26 in the Historical Data section shows how the rate of homeownership increases with age but then declines slightly after age 65. Again there is substantial variation with the under-35 age groups having much lower homeownership rates than the overall average and all groups over 45 having much higher homeownership rates.
• Married couples have higher homeownership rates than other family types. Eighty-four percent of married couples without children and 74 percent of married couples with children own a home. However, only 34 percent of single-parent households own a home.

For some groups homeownership is almost universal. For example,married couples who had children, who earned over $80,000, and whose head was 45þ54 years of age had a homeownership rate in 1991 of 97 percent. Even the minority rate for this group was 96 percent. Normally it is rare, even in the high-income groups, for the white and minority rates to be this similar. For some other groups, homeownership can be the exception. For example, only 14 percent of those households with a head under 25 years old own a home. This percentage increases to only 25 percent for married couples with children in the same age group.

The Federal Government encourages homeownership through favorable tax treatment and other incentives because homeownership is considered good for families and society. In particular, children are thought to benefit from growing up in owner-occupied homes. The post-1980 decline in homeownership rates has sparked public concern partly because the decline has been sharper for households containing children. In 1980, 71.0 percent of children lived in owner-occupied homes, compared to an overall homeownership rate of 65.6 percent. By 1991 the rate for children in owner-occupied homes, 62.6 percent, had fallen behind the homeownership rate for all households, 64.1 percent. The table above traces these rates from 1970 through 1991.

Spatial Patterns

The decennial census provides data on homeownership rates at various levels of geography.

In 1990 State rates varied from a low of 52.2 percent in New York to 74.1 percent in West Virginia. Fifteen States had homeownership rates more than 5 percentage points higher than the national average while five States had rates more than 5 percentage points lower than the national average. The States with the highest homeownership rates were predominantly the more rural States, such as Mississippi, Alabama, and Maine, but with notable exceptions, such as Michigan and Pennsylvania. New York, Hawaii, Nevada, California, and Alaska had the lowest homeownership rates.

The decline in the national average homeownership rate between 1980 and 1990 is reflected in the State data. Over this period 18 States experienced rising homeownership rates, but 32 States experienced declining rates. There appears to be no general pattern in the gains and losses. Both New York (the lowest rate State) and West Virginia (the highest rate State) had increases in their homeownership rates. Rural and industrial States had both increases and decreases as did States that gained or lost population over this period. Most of the changes were small; in fact, many of the changes were statistically insignificant. The largest increases in absolute terms were New York, 3.6 percentage points, and Maryland,3.0 percentage points. The largest decreases were Nevada, 4.8 percentage points, and Arizona, 4.1 percentage points.

Rural homeownership rates were substantially higher than urban homeownership rates in 1991, 80.6 percent versus 60.9 percent.Within urban areas central cities have the lowest homeownership rates, 48.6 percent in 1991 compared to 65.8 percent in the rest of the urban area.

One of the most interesting homeownership patterns is the variation of rates within central cities. Bogdon and Silver tabulated 1990 census data for 77 metropolitan areas, including the 50 largest metropolitan areas and at least one metropolitan area in each State.¹ The following table reports the homeownership rates for the 10 central cities with the lowest rates and the 10 central cities with the highest rates among the 77 metropolitan areas studied.

Cheyenne is approximately 20 percentage points higher than the average for all central cities, while Newark is approximately 17 percentage points lower than the average. There are a few outliers at the lower end; only the first five are more than 6 percentage points below the national average for central cities. But there are many more outliers at the upper end. With the exception of Tampa-St. Petersburg, San Jose, and Fort Lauderdale, the outliers at the upper end are typically central cities in smaller metropolitan areas. In general the larger central cities tend to have homeownership rates below or near the average for central cities.

Many things account for the wide variation among central cities.Central cities, such as Bergen-Passaic, that have a housing stock characterized by a large portion of rental units in buildings with more than one unit are less likely to have a high homeownership rate because conversion of existing rental units to ownership is difficult. Central cities with high housing costs, such as Los Angeles, are likely to have low homeownership rates. Where the ratio of central city population to metropolitan population is low,such as San Francisco, the central city homeownership rate is likely to be lower. Permissive annexation policies, which allow central cities to annex suburban growth, probably contribute to higher central city homeownership rates. Less dense central cities,such as Cheyenne or Boise, are likely to have higher homeownership rates. Rent control may be an explanatory factor for the low homeownership rates in New York City along with a large proportion of units in multiunit structures.

Consistent variation in homeownership rates can be found down to the census tract level. The following table presents data on homeownership rates for metropolitan census tracts disaggregated by median household income in the tract relative to local median income and by the percent of minority households in the tracts. The bottom line shows that, on average, homeownership decreases as the percentage of minority households in the tracts increase. There is only one break in this pattern. Similarly, the far right column shows that homeownership increases as tract income rises relative to local median household income. These results are perfectly consistent with the demographic information reported earlier.Income and race are primary determinants of homeownership rates.There are some interesting exceptions. When tract median income is less than 50 percent of area median income, there are many cases where homeownership rates actually rise as the percent of minority households rises. However, only a very small proportion of the population resides in these tracts. Tables for nonmetropolitan census tracts and all census tracts are similar except that there are more exceptions in the nonmetropolitan pattern.

Almost two-thirds of the metropolitan population resides in the census tracts in the box in the lower left-hand corner of the table. Even in this narrow range, there is once again a wide variation in homeownership rates, extending from 53.3 percent to 79.3 percent.

Dynamic Patterns

Movement and change characterize the American housing market.According to the 1991 AHS, 18 percent of all households had moved within the past year. Tenure choice frequently changes when a move occurs.

In a 1991 article in Housing Economics, David Crowe pointed out that, after 1980, the propensity of owners to remain owners after a move fell substantially.² He compared averages of AHS data for 1977 and 1979 with the same averages for 1987 and 1989. In the late 1970s, 67 percent of owners who moved owned their new dwelling unit. By the late 1980s, this proportion had declined sharply to 58 percent. Dr. Crowe updated his data for HUD using the 1991 AHS.These data showed a continuation of the late 1980s pattern with 57 percent of owners remaining owners after a move. Dr. Crowe also studied changes in the proportion of renters who become owners when they move. This proportion also declined between the end of the 1970s and the end of the 1980s, from 25 percent to 22 percent.

These changes play a significant role in the decline in homeownership over the 1980s. This can be illustrated by using data from the 1991 AHS on the number of moves by renters and owners. The moves recorded in 1991 resulted in an increase of approximately 400,000 in the number of owners. Approximately 1,700,000 households who had been renters became homeowners while another 1,300,000 households (2,940,000 minus 1,669,000) who had been owners became renters.

However, these moves would have generated a substantially greater increase in the number of homeowners if households had behaved in 1991 as they had in the late 1970s. Approximately 2,200,000 households who had been renters would have become homeowners while only 1,000,000 households (2,940,000 minus 1,982,000) who had been owners would have become renters. This would be a net increase of approximately 1,200,000 owners.

During a period of declining homeownership, one would expect the proportion of renters who become homeowners to diminish. The decline in the proportion of owners who remain owners after a move is more surprising. Dr. Crowe notes that this decline is consistent across all age groups. Although the AHS asks movers why they moved,HUDþs examination of these data reveals few clues to explain the decline. The reasons that one would expect to play an important part in the change from owner to renter status are cited by comparatively few movers. For example, less than 8 percent cite either change in marital status, lower expense, or other financial or employment reasons.

In his Housing Economics article, Dr. Crowe suggests some possible explanations. The 1986 Tax Reform Act reduced the incentives in the tax system favoring homeownership. Not only were marginal tax rates lowered, but the standard deduction was substantially increased.Combined, these tax law changes diminish the value of the mortgage interest and property tax deductions. Lower appreciation rates both reduce the wherewithal to purchase a new home and reduce the incentives to invest in ownership. While these factors certainly contribute to the decline, Dr. Crowe concludes that there is as yet no convincing explanation of why in recent years more owners who move are becoming renters.

Implications

There are wide demographic and spatial variations around the national average homeownership rate. The most troubling findings are the substantially lower homeownership rates of blacks and Hispanics, the decline in the proportion of children growing up in owner-occupied homes, and the substantial variation in homeownership rates across census tracts that can be related to income and race.

The patterns also suggest how homeownership fits into American life. The sharp differences in rates by income and age indicate that there is a life-cycle aspect to tenure choice. Preferences for homeownership may become stronger after age 25 or 30. Also the ability to afford a home increases with age both because of increased income and because of the longer period in which to acquire wealth. The variation in homeownership rates across cities suggests that population density and the type of housing stock affect the homeownership rate.

The dynamic patterns raise interesting questions about how tenure choice operates within the housing market. In particular less than 60 percent of all owners who move remain owners after the move. Unless a household were forced out of its home by financial problems, one would expect that an owner who moves could afford to remain an owner and has the sophistication to find a new home;therefore, the choice of rental status would appear to be a voluntary decision. The increasing propensity of owners who move to choose a rental option could signal a shift in preferences toward renting among American households. More needs to be learned about the reasons for this trend and its implications for homeownership. Nevertheless, it is clear from opinion surveys and the behavior of the large majority of families that ownership is still the preferred option.