Introduction

On June 17, 2004, in a speech to the National Federation of Independent Businesses in Washington D.C., President Bush stated “… if you own something, you have a vital stake in the future of our country. The more ownership there is in America, the more vitality there is in America, and the more people have a vital stake in the future of this country” (White House 2005). The underlying assumption of such statements is that most Americans can save and invest. In considering this assumption, a reasonable question to ask is, how do women fare in asset ownership and wealth accumulation, and what factors contribute toward higher asset levels for their families or households? After all, it is widely known women are more likely than men to live in poverty in the United States. Almost one-third of female householders have incomes below 100% of the poverty line and among Black and Hispanic female householders the percentage living in poverty approaches 40% (U.S. Census Bureau 2006a). According to U.S. Census data, the number of female householders has increased in recent decades from over five million in 1970 to 14 million in 2006 (U.S. Census Bureau 2006b). It thus stands to reason that women, much more than men, will find joining the Ownership Society a challenge.

Nearly 50 years ago, Titmuss (1962) argued that formulas of inequality in industrial societies must include examination of wealth as a demarcation of disparity. Sherraden (1991), in his seminal theoretical work on asset development, agreed with Titmuss, arguing wealth is a strong indicator of well-being. As Oliver and Shapiro (1995) stated, “Most people use income for day-to-day necessities; by contrast, assets often bring income, power, and independence” (p. 32). The risk of poverty for female householders is particularly of concern because gender income disparities translate into wealth disparities (O’Neill 2003; Schmidt and Sevak 2006). A lack of wealth affects individuals across the lifespan, while lack of income is more of a transitory condition (Oliver and Shapiro 1995; Wolff 2001). Thus, it is alarming that in 2001, female householders had average net wealth of $27,850, compared to $86,100 for all households in the United States (Consumer Federation of America 2004). In fact, female householders have the least amount of wealth of all family types in the United States and never married women with children have the lowest level of all. Men who head households have been shown to have about three times the average wealth of their female counterparts after controlling for income, personal, and labor market characteristics (Conley and Ryvicker 2003; Hao 1996).

While there has been a great deal of research examining economic well being among women, the vast majority focus on income and income-to-needs ratio measures. Wealth indicators have not been widely examined and research has paid little attention to wealth and asset accumulation among female householders specifically. Wealth, in addition to traditionally used measures of income, is an important determinant of economic status and life chances, especially later in life and for children.

This study aimed to descriptively document the wealth status of female householders and examine which factors are associated with female heads’ ability to accumulate assets. We hypothesized that characteristics of both the female head and the household will influence asset accumulation. Specifically we examined three questions: (a) Are there significant differences in asset patterns between female-headed households with and without children? (b) Does the presence of additional adults and their relationship to the household head influence asset accumulation among female householders? (c) Has asset ownership of female householders changed significantly since the mid-1990s when many states relaxed regulations on asset limits?

Theory of Saving

Savings rates and thus wealth accumulation will vary for a variety of reasons, including level of income, life cycle, and institutional structure. In neoclassical theory individuals are viewed as rational beings who seek to maximize individual satisfaction, largely a function of consumption. In the view of neoclassical economists, income and assets both represent resources that may be used to finance consumption. Individuals make choices weighing present and future consumption taking into account income and life cycle (Beverly et al. 2003; Friedman 1957; Modigliani and Ando 1957; Wakita et al. 2000; Yilmazer 2008). In order to smooth consumption over time, households must save during working years to finance consumption in retirement. Additionally, according to the life cycle hypothesis, consumption and saving will reflect where an individual is in their life cycle (Finke et al. 2006; Modigliani and Ando 1957). Younger households might be expected to have lower wealth due to lower earnings and more accrued debt; midlife households might be expected to save more for retirement and work on debt reduction; and upon retirement dis-saving begins to occur.

While neoclassical theory helps to explain savings outcomes, it does not fully explain wealth accumulation. Additionally, it is less applicable to poor households. Theorists observe that institutions matter in shaping and influencing opportunities, behaviors, and individual performance (e.g., Beverly and Sherraden 1999; Neal 1987; North 1990; Sherraden 1991). Institutions might be thought of as “purposefully created policies, programs, products, and services that shape opportunities, constraints and consequences” (Schreiner and Sherraden 2007, p. 30). According to this view, a great extent of saving is done through policies and structured programs; such as, the home mortgage tax deduction, 401(k) plans, and individual retirement accounts. Such programs largely benefit higher income people. The poor often do not participate in such policies, and benefits often come through the tax system that benefits the poor little if at all. Given the disproportionate number of female householders who live in poverty, they are much less likely to have the support of institutional structures.

Literature Review

Importance of Wealth

Research indicates that asset ownership and wealth accumulation have meaningful effects on the well-being of female householders and their children that go beyond income and have important intergenerational implications (Zahn and Sherraden 2003). Wealth has been shown to have positive effects on health (Robert and House 1996; Shea et al. 1996). Among female householders, wealth translates into increased child cognitive development, educational attainment and parental expectations for their children (Hao 1996; Zahn and Sherraden 2003). Household wealth is associated with increased self-esteem among adolescents (Axinn et al. 1997) and teenage girls whose parents have higher levels of wealth are less likely to become single teenage mothers (Conley 1999). Studies have also found that children whose parents accumulate wealth are more likely themselves to build wealth including homeownership; thereby, extending positive effects of wealth across generations (Henretta 1984; Oliver and Shapiro 1995; Pritchard et al. 1989). Further, owning assets changes one’s outlook on life, creating a more optimistic future-orientation and improving future economic opportunity (Shobe and Page-Adams 2001). Finally, assets represent potential security against future economic downturns or crises (Parks-Yancy et al. 2007; Sherraden 1991).

One of the primary forms of wealth in the United States is homeownership. Approximately one-third of wealth in America is in the form of owner-occupied housing (Wolff 2001). According to the 1999 American Housing Survey 81% of married couple households were homeowners, compared to 54% of single female householders and 42% of female householders with children (U.S. Department of Housing and Urban Development 2001).

As is the case with wealth in general, homeownership, in particular, is associated with a number of positive life outcomes. In families owning homes, children tend to have higher standardized test scores (Essen et al. 1978), are more likely to complete high school (Aaronson 2000; Kane 1994), have fewer non-marital pregnancies (Green and White 1997) and greater overall educational attainment (Scanlon and Page-Adams 2001). In addition, homeownership represents greater residential stability for families which may lead to increased economic opportunity across the lifespan (Rohe and Stewart 1996). Homeowners are four times less likely to move than renters (Stegman et al. 1995). Homeownership may also contribute to reduced periods of unemployment (Goss and Phillips 1997), elevated health status (Stronks et al. 1997), and increased likeliness of civic engagement (Rohe and Stegman 1994).

Unlike other forms of wealth, automobile ownership is not fiscally stable due to depreciation of investment across time. Nonetheless, it is a vital asset in the United States as it has been found to increase likelihood of economic independence by enabling families to access educational opportunities, employment, health care, child care, social supports, and community relationships (Brabo et al. 2003; Fletcher et al. 2005) and is, therefore, important in assessment of household financial well-being. In the United States a car represents both autonomy and social status for a family. While 89% of households own a car, 94% of welfare participants (most of whom are single women and their children) must rely on other people or public transit for transportation (Johnson 2000). Punctuating the importance of the ownership of a vehicle, Wilson (1996) makes the case that most employers who pay a living wage are prohibitively distant from low-income families who possess the skills and need, but not the transportation to take the job.

Finally, pensions play an important role in future economic well-being, especially for women who are more likely to experience poverty in old age. People in households where someone obtains income from a pension are provided some protection against poverty (McNamara 2007). While the ratio of women’s to men’s defined contribution plan accumulations increased from 40 to 44% between 1989 and 1998, it was concentrated among the age cohort 45–53. Among other age groups the gender pension gap increased between 1989 and 1998. And while most gender differences in defined contribution plan accumulation can be attributed to differences in earnings and job characteristics, men are significantly more likely to hold a pension (Employee Benefit Research Institute 2000). Additionally, liquid assets (e.g., short-term interest bearing assets such as savings accounts) are critical to smooth out short-term fluctuations in income and provide a safety net in the event of a financial crisis (Hong and Kao 1997; Xiao and Anderson 1997; Young and Hofferth 1998). A household is considered asset poor if their asset holdings are insufficient to meet basic needs (as measured by the income poverty line) for a period of 3 months (Haveman and Wolff 2004; Hong and Kao 1997).

Predictors of Wealth Accumulation

Much empirical attention has been given to wealth accumulation in the United States in the past decade. Studies have revealed the predictors of wealth include age, race, gender, educational attainment, family structure, and household income. Indeed being unmarried, minority race or ethnicity, and having low income and low education level greatly increases the odds of being in a bottom net worth quintile (Finke et al. 2006). Wolff (2001) found wealth accumulates through retirement age and then levels or moderately decreases. Other studies have shown glaring disparities between Whites and other racial/ethnic groups in both emergency funds and wealth attainment (Gittleman and Wolff 2004; Hong and Kao 1997; Keister 2004). In particular, African Americans lack wealth compared to their White counterparts. This is in part due to the advantage Whites have had in inheritance and historical dynamics in housing segregation and discrimination in lending against Blacks (Massey and Denton 1993). Given that African Americans make up a disproportionate number of female householders, race is an important factor to consider in examining wealth accumulation among female-headed households.

As noted, wealth also varies by gender. In most cases, controlling for differences in observable characteristics, female householders achieve lower levels of wealth compared to male-headed households (Conley and Ryvicker 2003). However, in a sub-sample of young households (age 26–39) the wealth differences between single females and single males disappear, suggesting either the wealth gap is changing among younger single generations or that wealth gaps emerge later in life among single heads of households (Schmidt and Sevak 2006). Yamokoski and Keister (2006) found strong evidence of a family gap in wealth with both single mothers and fathers compared to adults without children. The most severe discrepancy was among single mothers. The largest differences in wealth existed between female householders and married couples (Schmidt and Sevak 2006; Yamokoski and Keister 2006), with the greatest wealth levels found in married couple households in which both adults were employed (White and Rogers 2000). The number of employed adults and the number of children in a household also predicts ability to build wealth (Keister 2004).

Oliver and Shapiro (1995) demonstrated that educational achievement typically leads to jobs that pay high salaries, which, in turn, results in increased wealth. In general, women’s economic status declines following a divorce (Holden and Smock 1991). However, research indicates that among divorced women, human capital investments and receipt of child support have positive effects on women’s wealth after controlling for economic and personal characteristics of the household (Mckeever and Wolfinger 2001). Unfortunately, women’s educational and occupational statuses are negatively impacted by family responsibilities. Both married women and women with children are more likely to experience interruptions in schooling and employment (Groot et al. 1990; Sharpe and Baker 2007), causing wage penalties (Waldfoegel 1997, 1998) which in turn influence earnings and wealth accumulation.

A seemingly obvious correlation, households with higher incomes have more wealth on average. In part, households with high income levels amass more wealth because they are able to do so earlier, longer, and more aggressively than lower income households (Oliver and Shapiro 1995; Wakita et al. 2000; Wolff 2002).

While much is known about wealth disparities and the predictors of wealth accumulation in the United States, one area remains relatively unexplored. Comparisons of wealth disparities have been made between various racial/ethnic groups, age levels, household compositions, and genders. However, few studies have examined factors affecting wealth accumulation explicitly among female householders. Additionally, most wealth studies examine aggregate and net worth. While this study looks at total wealth and net worth, it is unique in its examination of different types of assets discretely. Given the benefits and long-term implications of different asset holdings, understanding the predictors of each has important implications for social policy.

Methods

Data

Data for this study were drawn from Wave 3 of the 1996 and 2004 panels of the Survey of Income and Program Participation (SIPP). The SIPP is a longitudinal survey conducted since 1984 by the U.S. Census Bureau. Each panel includes a nationally representative sample of adults, providing detailed information on household, family, and individual income and assets. The information is collected triennially for all individuals in the household for the 4 months preceding each interview.

Each wave includes a core questionnaire, asking basic sociodemographic information, and one or more topical module questionnaires, asking more detailed information on specific subjects. Wave 3 includes a topical module asking about financial assets. Data for the 1996 Wave 3 were collected between November 1996 and February 1997, and those for the 2004 Wave 3 were collected between November 2004 and February 2005. In each panel, the sample for this study includes one non-elderly (ages 25–64) respondent (known as the reference person) from each Wave 3 female householder. Households containing married couples were not included in the sample, though households may contain other, unmarried adults or dependent children (ages 0–17). The resulting data set includes information for 6,131 female householders in 2004 and 6,366 female householders in 1996. In each panel, nearly half of these households contained one or more dependent children.

In addition to separately analyzing data from the 2004 SIPP, a comparison was drawn with a similar sample of female householders drawn from Wave 3 of the 1996 SIPP. This comparison allowed a look at how asset holdings and values have changed for female-headed households over time.

Measurement

The dependent variables in this study are measures of asset holdings and values. Although it is common in research on wealth for all financial assets to be aggregated to the household level, we examined one asset (defined-contribution pension plans) at the individual owner level, specifically those pension assets owned by the female household head. All other financial assets were aggregated to the household level. We examined four asset categories: (a) defined-contribution pension plans, (b) non-pension interest-bearing assets, (c) home ownership, and (d) car or truck ownership. Defined-contribution pension plans are the sum of assets held in an individual retirement account (IRA) (SIPP variable TALRB); a 401k, 403b or thrift savings plan (SIPP variable TALTB); or a Keogh plan (SIPP variable TALKB). Interest-bearing assets include money held in interest-bearing checking, savings, money market, certificates of deposit (CDs), municipal or corporate bonds, or government securities (aggregated in SIPP variables THHINTBK and THHINTOT). Each asset category has two dependent variables measuring first, whether or not the asset was held, and second, the value of the asset for households who held these assets. For pension and interest-bearing accounts, it measured the dollar amount in these accounts in the past month. For home and vehicle ownership, the second dependent variable measures the value of equity owned.

The literature on asset holdings has identified several important control variables included in this analysis. These include (a) race, (b) ethnicity, (c) education level, (d) age, (e) employment status, (f) total household income and, for (g) home ownership, a measure of housing values in the region where the household resides. In addition we know that household structure and size influence income (Conley and Ryvicker 2003) and we hypothesized that these also influence the ability to save. Household structure was controlled for by variables indicating the presence or absence of children in the household and of extra adults in the household, specifically of a nonmarried male or female partner, a male or female relative, or a person aged 65 or older.

Race and ethnicity were both categorical variables with the impact of Hispanic/non-Hispanic ethnicity estimated separately from race. Education, age, and employment status were all measured categorically for the female household head. Total household income is aggregated within the Census Bureau and, for this study, was measured relative to the poverty level given household size and divided into four categories ranging from less than the federal poverty line to 300% of poverty or above. Because none of these female householders were married and their prior status may have had an impact on assets held, a categorical variable denoted those who were previously married (divorced or separated), or never married. Widows were omitted as preliminary analysis showed their asset holdings to be significantly different from those of other female householders. Dichotomous variables indicated receipt of child support or alimony and payment of fees for child care or care of a person with disabilities.

For estimation of the homeownership models only, four categories denoting the influence of geographic location on housing values were created based on median housing values in each state. States comprising each category are shown in Appendix Table 5. Years of home ownership and mobile home ownership were controlled for. For estimation of the defined-contribution pension, variables measuring the portfolio of investment products and the number of years the pension had been held were created.

Analyses

Data were analyzed both descriptively and in multivariate models. In addition, dependent variables measuring asset holdings and values were compared at two points in time to see whether change had occurred since the passage of legislation regarding assets that might affect low-income female householders (Sullivan 2006).

Because the value of each asset is conditional on that asset being held, there is a potential for selection bias. The potential non-randomness of the initial decision (or financial ability) to, for example, purchase a home creates the necessity to model these two equations as conditional, within the framework of a model that corrects for the selection bias, if it exists. Each of the asset categories were modeled using the Heckman selection model (Greene 1993), in which estimation of the second (value) equation was conditional on answering yes to the first (holding) equation. In each, the first (holding) equation was estimated using probit analysis. The second (value) equations were estimated using ordinary least squares regression. All estimation was completed in Stata allowing us to account for the complex survey design of the SIPP.

Because the two equations in each model were estimated sequentially, the set of independent variables in each must differ somewhat. In order to satisfy the mathematical demands associated with estimation of the Heckman procedure, we chose to allocate variables across the two equations where education level is thought to influence ownership of assets, but not values, and that household income level influences value of assets, but not ownership. Similarly, money into the household in the form of child support or alimony and money out of the household in the form of payments for care of a child or disabled household member were thought to influence asset values, but not ownership. Pension investment options, ownership of a mobile home, and years held (for both pensions and homes) were also thought to influence asset values, but not ownership.

Results

Descriptive Statistics

A weighted profile of the characteristics of female householders included in this study is shown in Table 1. Significant differences existed between households with one adult (the female head) and those with multiple adults, as well as between households with and without children. The householders with one adult were, on average, younger. They were less likely to be of Hispanic ethnicity, to have a work-limiting disability, or to have preschool age children than householders who live with more than one adult. These one-adult heads had higher levels of completed education and were more likely to be never married, but had lower total monthly household incomes and were more likely to have annualized incomes that left their households in poverty than multiple-adult householders.

Table 1 Weighted characteristics of the sample (n = 6,131)
Full size table

Within both one-adult and multiple-adult households, those with children were 5–10 years younger than those without, on average, and were more likely to be African American and/or of Hispanic ethnicity. These women were more likely to have just completed high school than women without children and were less likely to have four or more years of college. They were less likely to work full time, more likely to be previously married, and more likely to have incomes below the poverty line. Nearly half of one-adult households with children received monthly child support payments compared with just over one-third of multiple-adult households with children.

Multiple-adult households were more likely to own their home and one or more vehicles, while in one-adult households the household head was more likely to have a defined contribution pension or to own interest-bearing assets (Table 1). Households without children were significantly more likely to own assets in all four asset categories compared with households with children and, in each category, the value of assets owned was significantly higher in households without children. Total household wealth and net worth were lower in both one-adult households and households with children. Even for multiple-adult female households, however, asset ownership levels were low compared with overall statistics for the United States. Only half of the multiple-adult female households in the SIPP owned their own home (44% for one-adult households) and fewer than half were contributing toward a defined-contribution pension.

Multivariate Analyses

Factors associated with ownership and value of interest-bearing assets and defined-contribution pensions by household heads are shown in Table 2. Tables 2 and 3 report both the probability of ownership (derived from the probit selection equation in the Heckman model) and the marginal effects on value for those who owned these assets, simplifying interpretation of the relative magnitude of the effect of each independent variable.

Table 2 Probability of ownership and marginal effects on equity values from Heckman selection model of factors affecting household-owned interest-bearing assets and defined-contribution pensions owned solely by the household head
Full size table
Table 3 Probability of ownership and marginal effects on equity values from Heckman selection model of factors affecting household home and vehicle ownership and equity
Full size table

Interest-Bearing Assets

Descriptive statistics (Table 1) show that households without children were most likely to own interest-bearing assets. The value of interest-bearing assets was lower in households with children, but overall higher in one-adult households than in multiple-adult households. Households with older heads (ages 55–64) who were White, worked full time, had at least some post-high school education or training were more likely to own interest-bearing assets (Table 2). Those with higher annual incomes and/or had more education held higher values of interest-bearing assets, as did those who had never been married and, in multiple-adult households, those who lived with a female partner. While the coefficients on employment of the household head and values of interest-bearing assets seem counterintuitive, it is likely this relationship is associative, not predictive. That is, while female heads who work full time are more likely to own interest-bearing assets, when those who work less than full time do own interest-bearing assets they are of higher value. These households may receive much of their income from interest-bearing assets, allowing them to work less. Ownership of these high value interest-bearing assets may be that of the female head or another adult household member. More than half of these households contained an adult with a work-preventing disability so they may receive disability income as well.

When other demographic and socioeconomic characteristics of the household were controlled for (see Table 2), both extra adults and children still exerted influence on holdings and values of interest-bearing assets. In both one-adult and multiple-adult households, children negatively impacted the likelihood of holding assets in interest-bearing accounts yet exerted no significant influence on the value of these assets. In multiple-adult households, interest-bearing assets were nearly $9,500 higher if the extra adult was a female partner, holding all else constant.

Defined-Contribution Pensions

Pension holdings were more likely in households with an older, White, well-educated head who was employed full time (Table 2). Children exerted a negative influence on pension ownership. A male or female partner in the household exerted a positive influence on pension ownership and, if a male partner, significantly increased the value of the pension. Pension account values were higher among heads who lived alone and had never been married. Values increased with both the education level of the household head and with household income level. Pension values were significantly higher in one-adult households for those investing in stocks, government or corporate bonds, or government securities. In multiple-adult households, pension values were highest for those investing in stocks, money market accounts, and CDs. In both household types, pension values increased by $3,700 to $4,500 for every year they were held.

Home Ownership

Descriptive statistics (Table 1) show lower home ownership among one-adult households and households with children. Home equity was significantly higher for households without children, yet there was no significant difference in home equity in one-adult or multiple-adult households overall. Once demographic and socioeconomic differences are controlled for (Table 3), the influence of children disappeared. Marital status of the household head was important with never married heads less likely to own their home. In multiple-adult households, household heads with a female partner were more likely to own a home, but have significantly lower home equity (nearly $12,000 lower) than households with other adult relationships. Home ownership was less likely among householders who were younger, minority, Hispanic, and had lower levels of education. Home equity was lower in one-adult households with younger heads, in African American households, and in multiple-adult Asian households. Although householders who worked full time were more likely to own a home, home equity was significantly higher among householders who worked either part time or not at all. This dynamic is likely similar to that seen in interest-bearing asset holdings. Perhaps the largest influence on home equity is the type of home owned as those who owned mobile homes had 65–70% lower equity than those owning conventional homes. Relative real estate values matter significantly with all households less likely to own homes in states with the highest median housing values. For those who were able to purchase a home in these markets, however, home equity was, on average, about three times higher than for those who owned a home in one of the states with the lowest median housing values.

Vehicle Ownership

Cars and trucks are the assets that female householders are most likely to own. Still, both vehicle ownership and equity in vehicles owned was significantly lower in one-adult households and households with children (Table 1). Households with multiple adults own two vehicles on average, while households with one adult own one. Vehicle ownership patterns (Table 3) are similar to those seen for other household assets with older, White, previously married, better educated and employed household heads more likely to live in households owning one or more vehicles. Among minorities, African–American households were least likely to own a vehicle and had less equity in the vehicles they owned than other racial and ethnic groups. Multiple-adult households were more likely to own a vehicle if one of the adults in the household is the male partner of the household head. A female partner in a multiple-adult household reduced the probability of owning a vehicle, but had no impact on vehicle equity. As education level of the household head increased so did the amount of vehicle equity in both one-adult and multiple-adult households.

Change Since 1996

The top section of Table 4 shows the percentage of female householders in 1996 and 2004 holding assets in each of the categories examined in this paper. These are cross-sectional samples and, thus, are households with the same basic characteristics at each point in time rather than the same households interviewed at two points in time. The percentage of female householders who owned assets increased significantly in nearly all categories (weighted t-tests).

Table 4 Assets held by female householders, 1996 and 2004
Full size table

The distribution of relative asset values is shown in the bottom section of Table 4. Because the value of the dollar changes each year with the increase in inflation, asset values for 1996 cannot be directly compared with those for 2004. Rather than simply using the Consumer Price Index to convert 1996–2004 dollars, the values of each asset were divided by a base value to create a relative value for each asset that could be compared across the two time periods. For the assets held by the household head, interest-bearing assets and defined contribution pensions, this base value was the total income for the female householder last month. Relative values of these assets, thus, are shown as either a fraction or a multiple of monthly income. For example, in 1996 the median one-adult female householder with children had interest-bearing assets equivalent to 40% of their monthly income while defined contribution pension funds were over two times their monthly income. By 2004, the median one-adult female householder with children had interest-bearing assets equivalent to only 23% of their monthly income while defined contribution pension funds were about three times their monthly income. Similar results are seen in multiple-adult households. Median is used in this table rather than mean due to skewness inherent in asset values.

Each of the assets held by the household, i.e., home and vehicle ownership, were examined relative to their market values. Home equity was examined relative to total property value and vehicle equity was examined relative to total vehicle value. For example, in 1996 the median multiple-adult household without children had equity in their house equivalent to about 67% of its total value and owned their cars outright (100% equity). In 2004 households fitting the same description had 55% equity in their homes and 45% equity in the vehicles they owned. Similar declines in equity values are seen for other household types, both with and without children. Despite these declines, for most household types both net worth and total wealth have risen significantly relative to total monthly household incomes (bottom of Table 4). For example, in 1996 the median female householder had net worth above 3.5 times their monthly income. This rose to nearly five times their monthly income in 2004. Similarly, total wealth rose from nearly five times monthly household income to nearly seven times monthly household income. Still, net worth and total wealth holdings were considerably lower in households with children than in those without children.

Discussion and Implications

This study posed three research questions. First, are there significant differences in asset patterns between female householders with and without children? Descriptive analyses indicate that female heads with children, in both one-adult and multiple-adult households, were less likely than households without children to own interest-bearing assets, defined contribution pensions, a home, or vehicle. Controlling for personal and socioeconomic characteristics, the likelihood of owning interest-bearing assets or a defined contribution pension was significantly reduced by the presence of a child in the household. However, when a woman did own these assets, children had no negative impact on equity. This suggests that the barrier may be at the entry point, getting an account started, rather than an inability to add to the account balance over time. Savings mechanisms; such as, the Child Trust Fund recently adopted in the United Kingdom (Gregory and Drakeford 2006), may provide a model that could be adapted for use in the U. S. Interestingly, the presence of a child had no impact on home or vehicle ownership or equity.

Second, does the presence of additional adults and their relationship to the household influence asset accumulation in female householders? The impact of an extra adult in the household is also somewhat mixed depending on asset. Household heads with a female partner were more likely to be homeowners and to have a defined contribution pension plan. These same female partners reduced the probability of owning a vehicle, while a male partner significantly increased the probability of owning a vehicle. Female household heads with male partners were less likely to own defined contribution pension plans, but those who did had significantly more money in them. Surprisingly, sharing a house with a relative appears to have little, if any, significant impact on asset ownership or values.

Third, has asset ownership of female householders changed significantly since the mid-1990s when many states implemented reforms in asset limitations? Between 1996 and 2004, the percentage of ownership among female heads overall increased significantly in all four asset categories. While the value of interest-bearing assets to income ratio declined between 1996 and 2004, the ratio of defined contribution value to income increased significantly between 1996 and 2004. This increase may signal forced defined contribution participation by employers who are getting out of the defined benefit pension business. It also may suggest that some substitution of savings mechanisms may be taking place over time, with households choosing to invest in pension funds rather than leave their money in more readily accessible savings accounts. Interestingly, between 1996 and 2004 equity values in homes and vehicles declined significantly. This could suggest that due to a healthy economy, including lower interest rates and readily available credit, women were purchasing more expensive cars and homes. It might also indicate that due to the number of women moving from welfare to work, more low-income women purchased cars and homes thus lowering overall equity ratios. While ratios for home and vehicle ownership declined, ratios of household net worth and total wealth to household income increased significantly between 1996 and 2004, revealing that overall female householders are doing better in wealth accumulation.

After controlling for demographics and socioeconomic factors, in general female householders who were older, White, better educated, and worked full time were more likely to own assets. Overall, characteristics and socioeconomic factors appear to have a greater role in whether or not a female head owns an asset than its value.

Previous research has provided evidence of the positive effects assets can have on individual and household well-being. Thus, many women are missing out on these potential benefits. In particular adult female householders with children, non-White, lower-income women, younger women, and less educated women are at a disadvantage in ownership and equity of assets. Additionally, structural issues including women’s disadvantage in the labor market through lower earnings and less opportunity for mobility are likely to play a role in women’s ability to accumulate assets. While the Earned Income Tax Credit (EITC) and initiatives; such as, Individual Development Accounts (IDAs), may help structure access to asset ownership, other efforts are clearly needed. Although the proportion of women who have pensions has risen sharply over this period, less than half the female householders in this study in 2004 are defined contribution pension holders. This indicates that either women are not taking advantage of pensions when available from their employers or are not provided the opportunity to invest in defined contribution pensions. One remedy shown to increase participation is for employers to require employees to opt out of participation in these savings plans rather than the more common requirement to opt in (O’Neill 2007). While slightly more than half of female households hold interest-bearing assets, those who don’t are likely to be more susceptible to income shocks and perhaps more likely to turn to public assistance support without a safety net (Young and Hofferth 1998). Pensions and homeownership in particular have important implications for the economic well-being of older women. The added income of a pension provides some protection against poverty and owning a home reduces monthly housing costs and provides a potential source of added income; such as, through a reverse mortgage.

In addition to policy support, findings of this study point toward programmatic response at the community level and heightened awareness among financial service providers and such professions as social work. Part of the role of social workers working with low-income women, for example, should be to facilitate connection to saving and asset accumulation opportunities. Connecting women to banking services, economic education classes, and asset development programs; such as, homeownership programs and IDAs are examples (Garasky et al. 2008; Han and Sherraden 2009; Haynes-Bordas et al. 2008).

Limitations

This study provides an initial look at asset holdings among female householders. While the SIPP is a rich source of information about asset ownership and value, it lacks policy-relevant variables that may better explain why female householders tend to have lower assets than households headed by males. For example, we do not know whether or not women in this study have access to the institutional structures (low-cost checking or savings accounts) that are critical to the accumulation of interest-bearing assets and that may lead to banking relationships that help facilitate investment in longer-term assets such as IRAs and homes. Perhaps more importantly we do not know why women do not invest more. Is it because (a) their incomes are lower on average, than those in male-headed households, (b) the presence of a child lowers their risk tolerance (Chaulk et al. 2003), or is it (c) women prefer to better understand investment options and strategies before engaging in those markets? The poor have been shown to save when provided institutional structure and opportunity such as through IDAs. Indeed in one national demonstration of IDAs, low-income women were more likely to be savers than men (Schreiner and Sherraden 2007), suggesting that income levels, though important, are only part of the explanation. Future studies that examine both personal and household characteristics as well as institutional structures are needed to more fully understand the wealth accumulation capacity of women.