The November 2001 issue of U.S. Housing Market Conditions used American Housing Survey (AHS) data to present a picture of the first-time homebuyer market by showing trends and market shares for first-time homebuyers as well as their characteristics. This issue will continue to investigate characteristics of homebuyers by focusing on marital status and gender.

Historically, homeownership rates have been highest for households headed by married couples. The 1999 AHS, a national survey of the housing market funded by the U.S. Department of Housing and Urban Development and conducted by the U.S. Bureau of the Census, shows an 81.1-percent ownership rate for married couples, compared with 33.8 percent for households headed by separated individuals and 51.6 percent for households headed by singles (never married, divorced, or widowed). (See table 1.) Despite these persistent disparities, the following analysis will show that the 1990s were years of increased lending to nontraditional households. Lenders introduced special programs for first-time homebuyers and revamped their underwriting standards to consider the special circumstances of low-income families. Private mortgage insurers, the Federal Housing Administration (FHA), Fannie Mae, and Freddie Mac also played important roles in what some observers called "a revolution in affordable lending." Their efforts also benefited nontraditional, particularly female-headed, families. Insights about these homebuyers can be gained from the AHS.

The next section briefly describes the AHS. Following this, data on ownership patterns and home purchases by marital status and gender are reported for the years 1991-99. The last section discusses the role of FHA in the homebuyer market.

The American Housing Survey

The AHS began in 1973, and the national sample¹ now collects data in each odd-numbered year. The most recent available survey (1999) sampled 58,400 housing units, a sampling rate of approximately 1 per 2,000. The AHS uses computer-assisted personal interviewing techniques to gather information on all aspects of housing: alterations and repairs, costs, equipment, household composition and income, mortgages, neighborhood characteristics, recent movers, and structure and unit characteristics. Data from the AHS are published as tabular reports and as public-use microdata, where each record of the dataset corresponds to one housing unit. This article presents special tabulations of the microdata from the 1993, 1995, 1997, and 1999 surveys.

This article presents estimates of owner households who purchased units by year, marital status, gender, and first-time buyer status. Households were classified into three marital status groups: married, separated, and single. The "single" category included households in which the householder was never married, divorced, or widowed. Of course, the subsample of owners who move during the period between AHS surveys is much smaller than the total AHS sample. For example, although the 1999 survey sampled 58,400 units, only 6,040 of these were purchased between 1997 and 1999, and, of these, 2,394 were purchased by first-time buyers. The previous survey years provide samples of approximately the same size as the 1999 survey. Thus, our analysis is based on approximately 2,000 total and 850 first-time buyer observations per year. Although these sample sizes are adequate for generating statistically reliable totals, care must be taken when subdividing this sample further, such as by marital status or other household characteristics.²

Home Purchases by Single Women, 1991-99

Homeownership Rates. In the 1999 AHS data, husband-wife households comprised 53.8 percent of all households, compared with 3.1 percent for households headed by separated individuals and 43.1 percent for households headed by single individuals. (See table 1.) Households headed by singles increased only slightly from their 42-percent share of households in the 1993 survey. Although overall ownership rates increased from 64.7 percent to 66.9 percent from 1993 to 1999, ownership rates increased for married couples from 79.4 percent to 81.1 percent, and single heads of household from 48.5 to 51.6 percent. Although households headed by separated individuals had a slight drop in ownership rates—from 35.4 percent to 33.8 percent—they represent a small proportion of the population and are in a period of economic change as they transition from married to single status. Therefore, the remainder of the discussion will focus on a comparison of households headed by married couples and those headed by singles.

In addition to the disparity in ownership rates by marital status, homeownership rates differed with the presence of children in the household. Within each marital status group, households with children had lower ownership rates than households without children. In 1999, married couples had a 77.3-percent ownership rate if they had children and an 84.6-percent rate if they did not. Single-headed households with children had a 44.2-percent ownership rate compared with a 53.7-percent rate for single-headed households without children. Note, however, that if marital status is ignored and if the homeownership rate of households with children (67.1 percent) is compared with the rate for households without children (66.8 percent), the pattern is reversed. This is because married couples, who have higher overall homeownership rates, comprise the majority of households with children (25.7 percentage points of the total 36.5-percent share of all households) and the minority of households without children (28.1 percentage points of an overall 63.5-percent share of households).

The pattern was more complex when households headed by single males were compared with households headed by single females. Among male-headed households, the ownership rate was higher for those with children (51.2 versus 46.6 percent). Female-headed households, comprising more than 60 percent of noncouple households, conformed to the overall pattern; that is, those with children had a 41.9-percent ownership rate, whereas those without children had a 58.9-percent rate. (See table 1.) Except for male heads of household with children, whose ownership rate dropped from 54.3 percent to 51.2 percent, every one of these groups had a higher homeownership rate in 1999 than in 1993. In fact, female-headed households with children had the largest increase in homeownership rates of any group from 1993 to 1999, increasing their rate by 4.4 percentage points, compared with 2.3 percentage points for the overall population. The increase was particularly large for minority female-headed households with children. African-American female-headed households with children increased their ownership rate by 6.2 percentage points, whereas Hispanic female-headed households with children increased their rate by 8.2 percentage points.

Home Purchases. Analysis of AHS data shows that nearly 31 percent (13 million) of the 44 million home purchases in the 1990s were made by single-headed households. Forty percent of total home purchases (17 million) were made by households that were becoming homeowners for the first time, and households headed by singles represented 36 percent of these purchases. As discussed in the November 2001 issue of U.S. Housing Market Conditions, the overall trend in home-purchase volume in the 1990s was positive. Total purchases in 1999 of 5.6 million homes were higher than for any other year in the 1990s.

Single-headed households represented a slightly rising share of all homebuyers over the decade. In 1991-93, single-male-headed households comprised 12.5 percent of all homebuyers, compared with 14.7 percent in 1997-99, whereas single-female-headed households increased modestly from 16.8 percent to 17.7 percent of all homebuyers. (See table 2.) Growth was evident in households both with and without children, but growth was strongest for single-male-headed households without children and for single-female-headed households with children (9.9 to 12.1 percent and 4.9 to 6 percent, respectively). The small change in the share of single female buyers masks the fact that White female homebuyers actually declined slightly from 13.6 to 13.4 percent of the total, whereas minority female homebuyers rose from 3.2 percent to 4.3 percent. Both White and minority single male homebuyers increased slightly as a share of the total—from 10.6 percent to 12.0 percent for Whites and from 1.9 to 2.7 percent for minority buyers.

First-Time Homebuyers. Given their lower ownership rates, it is not surprising that a larger share of single-headed household purchases were made by first-time buyers. In addition, the proportion grew over the decade. During 1991-93, 48.3 percent of single-male homebuyers and 37.5 percent of single female homebuyers were first-time buyers, compared with 37.5 percent of all homebuyers. In 1997-99, the proportions increased to 52.7 percent for single-male-headed households and to 45.2 percent for single-female-headed households, compared with 40.1 percent for all homebuyers. (See table 2.) The rising percentages showed growing homeownership opportunities for single heads of households both with and without children. In the earlier part of the decade, single-female-headed households were no more likely than the average household to be first-time buyers, but by the end of the decade, single-female-headed households were more likely than average to be first-time buyers. Increases were particularly dramatic for minority single heads of households with children. During 1991-93, 60.2 percent of single-male-headed and 59.6 percent of single-female-headed minority households with children were first-time homebuyers. In contrast, the 1997-99 first-time shares were 73.6 percent for single-male-headed and 69.8 percent for single-female-headed homebuyer households with children.

Low Incomes and Homeownership. The most obvious barrier to homeownership by single heads of household is income. In 1999, single heads of households (both owners and renters) had incomes less than half those of married couples. Whereas married households had a median income of $52,000, single households' income was only $24,000. (See table 3.) Furthermore, the income gap between single-male- and female-headed households was substantial. Single-male-headed households had a median income of $30,516, compared with $20,000 for single-female-headed households. Among all household types that had achieved homeownership by 1999, those with children had higher incomes than those without children. Because children impose costs on the household, it is not surprising that the income threshold for ownership may be higher if children are present. Married homeowners with children had a median income of $65,000 compared with $50,000 for those without children. Similar patterns were observed for single heads of household: single male owners with children had a median income of $41,500, compared with $36,500 for those without children, and single female owners with children had a median of $29,000, compared with $21,000 for those without children. (See table 3.)

Not surprisingly, first-time homebuyers tended to have lower incomes than existing homeowners. On average, first-time homebuyers in 1999 had a median income of $42,000 compared with $45,400 for those who were already homeowners. (See table 3.) This pattern held for married households with children and for single-male-headed households with and without children. However, married first-time homebuyers without children, as well as single female first-time buyers either with or without children, had higher incomes than those of existing homeowners. For married homeowners without children, this pattern may simply reflect a greater likelihood of higher income households to become homeowners before they have children. There are at least two potential contributing factors to the observed pattern for single females. First, existing single female homeowners may have become homeowners when they were married and had a higher household income. Second, single females wishing to become homeowners may face additional financial barriers to ownership beyond income to support the monthly payments, such as expectations that repairs and maintenance will have to be hired out.

Summary. Ownership rates vary substantially by marital status and presence of children. Married couples have a much higher homeownership rate, whereas the presence of children in all household types is associated with a lower ownership rate. Nevertheless, among all household categories, owners with children had higher incomes than owners without children, suggesting a higher income threshold to achieve ownership if children are part of the household.

Single females with children are less likely than other singles to be homeowners. Although evidence shows that single-female-headed households with children have made important strides in homeownership in the 1990s, their low incomes will continue to be a barrier to achieving homeownership. In 1999, renter households headed by single females with children had a median income of only $16,480, barely more than half the $31,000 median of such households that achieved homeownership in that year.

FHA Loans by Gender and Marital Status

As explained in the November 2001 issue of U.S. Housing Market Conditions, FHA has long been recognized as the mechanism for young, low-income, and minority families to purchase a home. Families and individuals with little cash are attracted to FHA's low downpayment requirements, liberal income qualification guidelines, and flexible credit standards. FHA provides a unique mortgage product for borrowers who have found obtaining mortgage financing in the private conventional market difficult. The November 2001 U.S. Housing Market Conditions article shows that FHA played an important role supporting the homebuying market during the 1990s. In total, FHA insured over 6 million home purchase loans between 1991 and 2000. The share of FHA-insured home purchase loans going to first-time buyers increased from 62 percent in 1991 to 81 percent in 2000. The article also shows that FHA played a major role in the minority homebuying market as African-American and Hispanic borrowers, in particular, relied heavily on FHA loans during the 1990s. The share of FHA's home purchase loans going to African-American and Hispanic borrowers increased from 19 percent in 1991 to 34 percent in 2000.

This section expands on the earlier analysis of FHA—first by using FHA's own data to examine gender and marital characteristics of those obtaining FHA-insured loans, and then by using Home Mortgage Disclosure Act (HMDA) data to examine the share of the market for single borrowers captured by FHA. Note that the definition of "single" has changed to mean "unmarried" for the following discussion.

Characteristics of FHA Loans. Between 1993 and 2000, single borrowers accounted for 45 percent of FHA-insured home purchase loans, with single female borrowers accounting for 21 percent of FHA loans and single male borrowers accounting for the remaining 24 percent.³ The single-borrower share of FHA loans was 41 percent in 1993, rose to 47 percent by 1998, and remained there during 1999 and 2000. The share for single females increased from 19 percent in 1993 to 23 percent in 1998-2000, whereas the share for single males increased from 22 to 24 percent. Single borrowers accounted for a slightly higher portion of FHA's first-time borrowers than they did of FHA's repeat borrowers. For example, between 1997 and 2000, single female borrowers accounted for 23 percent of FHA loans for first-time homebuyers and 20 percent of FHA loans for repeat homebuyers; the corresponding percentages for single male borrowers were 26 and 19 percent, respectively.

The racial and ethnic distribution of FHA-insured home loans originated between 1993 and 2000 was as follows: 12.8 percent for African-American borrowers, 15.5 percent for Hispanic borrowers, and 71.8 percent for remaining (mostly White) borrowers. When FHA's single borrowers are considered as a separate group, African-American borrowers accounted for a slightly higher share (16 percent) than their overall distribution within FHA, whereas Hispanic and remaining borrowers (13.1 and 70.9 percent, respectively) accounted for slightly lower shares of FHA's single borrowers. In fact, single borrowers accounted for more than half (56.7 percent) of FHA loans going to African-American borrowers, and single female borrowers accounted for most of these loans, representing 36.9 percent of FHA loans going to African-American borrowers during the 1993-2000 period. The share of all FHA loans going to single female African-American borrowers increased from 3.3 percent in 1993 to 5.6 percent in 2000. As the next section shows, FHA has played a major role in the overall market for African-American households headed by single females.

FHA's Market Shares. The remainder of this section examines FHA's role in the overall mortgage market for single borrowers. HMDA data for metropolitan areas are used here because these data include all mortgage originations, as well as loans insured by FHA, thus allowing computation of market shares for FHA. HMDA does not identify single borrowers but does report the gender of both primary borrower and coborrower. This market analysis uses a proxy for loans originated by single borrowers, defined as loans for which the coborrower is reported by lenders as not applicable (rather than as male, female, or missing). Following this proxy procedure provided similar estimates of FHA's single borrowers as those reported above based on FHA's own data—both report that 45 percent of FHA's home purchase loans between 1993 and 2000 were for single borrowers.⁴

Based on HMDA data, FHA accounted for 22.3 percent of loans for single borrowers between 1993 and 2000, which was only slightly higher than FHA's share (18.9 percent) of all home loans originated during this period.⁵ As reported earlier, single borrowers accounted for an increasing share of FHA's business during the 1990s. However, this did not translate into an increasing FHA share of the mortgage market for single borrowers, as the non-FHA single market was also increasing during this period. FHA's share of the mortgage market for single borrowers was practically the same in 1993-95 (22.4 percent) as it was in 1997-2000 (22.1 percent). FHA's share of the single-female market was 24 percent throughout the 1990s, which was modestly higher than its 21.0-percent share of the single-male market.

FHA's share of the single-borrower market is larger when the analysis focuses on African-American and Hispanic borrowers. FHA accounted for 41.2 percent of all loans to single African-American borrowers between 1993 and 2000—more than twice FHA's share (19.2 percent) of the home purchase market for single White borrowers.⁶ FHA's share of the market for single African-American borrowers also increased during this period from 39.1 percent during 1993-95 to 41.8 percent during 1997-2000. Female borrowers have accounted for almost two-thirds of FHA-insured loans going to single African-American borrowers. FHA's share of the market for single female African-American borrowers averaged 45.1 percent between 1993 and 2000; the FHA share of this market has also increased, rising from 42.8 percent during 1993-95 to 45.8 percent during 1997-2000.

FHA's share of the market for single Hispanic borrowers rose from 31.7 percent during 1993-95 to 38.3 percent during 1997-2000, demonstrating FHA's increasing role in the Hispanic market. Male borrowers have accounted for two-thirds of all single Hispanic borrowers receiving FHA insurance, the reverse of single African-American borrowers (see above). It should also be noted that FHA has played an even larger role in the market for married Hispanic borrowers, accounting for 43.5 percent of the mortgages originated for these borrowers during 1997-2000.

Notes

1. The AHS also includes a metropolitan sample, but that was not used in this article.

2. See "First-Time Homebuyers: Trends From the American Housing Survey," U.S. Housing Market Conditions, 3rd Quarter 2001, U.S. Department of Housing and Urban Development, Office of Policy Development and Research, November 2001 for more detail about the AHS.

3. FHA data reported in the remainder of this section exclude data from Puerto Rico, Guam, and the Virgin Islands, which means that Hispanic borrowers are slightly underestimated in this analysis. The analysis starts with the year 1993 because that is the first year of complete marital status reporting in FHA's database. FHA classifies borrowers as married, not married (i.e., single), separated, or missing marital status. After 1993, the latter two categories accounted for less than 1.5 percent of FHA's home purchase loans.

4. Both sources also report that single females accounted for 21 percent of FHA's business between 1993 and 2000. In some cases, the correspondence between HMDA and FHA data is not always exact; for example, FHA reports that single African-American borrowers accounted for 7.2 percent of its home loans between 1993 and 2000, whereas HMDA reports that single African-American borrowers accounted for 8.8 percent of FHA-insured home loans during this period.

5. HMDA covers most, but not all, of the mortgage market in metropolitan areas. To the extent that HMDA's coverage of FHA loans is better (worse) than its coverage of non-FHA loans, FHA market percentages reported in the text will overstate (understate) FHA's market share. It should also be noted that missing gender, race, and ethnicity data are included in the total volume figures for FHA and the market but are not reallocated among the various gender, race, and ethnicity subgroups. The market share analysis begins in 1993 because prior to that year certain mortgage bankers were not required to report any HMDA data.

6. It should be noted that FHA's market share for single African-American borrowers (41.2 percent) was 7.3 percentage points higher than its share (33.9 percent) for married African-American borrowers. Thus, while FHA provides important support for married African-American borrowers, it provides even greater support for single African-American borrowers.