1 Gender and Wealth

Research interest in the super rich has grown dramatically recently as it becomes clear that these elite households own vast resources and control important political, economic, and social institutions that affect all households (Kelly & Volscho, 2014; Khan, 2012; Rivera & Tilcsik, 2016; Volscho & Kelly, 2012). Wealth is a particularly important part of the conversation because assets provide additional security beyond the benefits of income (e.g., a buffer against financial emergencies) and because wealth can create more wealth and be passed to future generations creating family dynasties. Despite growing research interest in top wealth holders, most work on the super rich has been gender-blind or at least gender-neutral. The gender-neutral approach in studies of high-wealth households, in part, reflects conceptual and analytic challenges in studying gender and wealth when most married couples hold assets jointly. Couples tend to merge their financial assets, including checking and savings accounts, the most common financial vehicle owned by households. Similarly, the majority of American couples buy a house at some point, and they tend to buy that residence jointly. Couples also tend to hold other assets together, including business assets, vacation homes, and the like. Retirement accounts (including Individual Retirement Accounts and pooled investments that accrue through employers) and cash accounts (e.g., checking accounts, savings accounts, Certificates of Deposit) are exceptions and can be owned separately by members of a couple. Importantly, however, most couples treat even separate bank accounts as joint property and pool resources held in these accounts to save for shared goals and to pay for consumption needs (Hamplova & Bourdais, 2009). The reality of shared property ownership becomes clear in the event of a divorce: when a couple divorces, the courts treat property, including wealth, as joint property; indeed, there is explicit recognition that the principles of shared assets and joint contribution promote “gender equality and purposefully equalizes the treatment of market and home labor” (Kelly, 2004, p. 208).

Previous research acknowledges the challenges associated with studying gender and wealth and offers some ideas about how to isolate and understand the role that gender plays in asset ownership and accumulation. One strategy is to focus on the assets women can own individually such as retirement accounts (Chang, 2010; Edlund & Kopczuk, 2009); this approach provides some insight into women’s wealth holdings, but it ignores the notion of joint ownership and treats assets as if they are not shared. This approach also leaves open questions about assets that tend to be owned jointly (e.g., the home) unless those assets are divided — for empirical purposes — between spouses. Again, however, this approach ignores the reality that most couples treat assets as shared property. An alternative conceptual and analytic strategy — the strategy we use in this paper — is to acknowledge that couples own property together, to avoid trying to attribute ownership to individuals, and to focus empirically on three groups: unmarried women, unmarried men, and married couples (Chang, 2010; Yamokoski & Keister, 2006). We opt to use this strategy because it allows us to isolate gender for unmarried people and to compare the unmarried to couples, who tend to operate financially as a unit that contains, in the case of different-sex couples (our focus) both a man and a woman.

2 The Wealth of Super-Rich Women, Men, and Couples

Four interrelated processes are likely to create differences in the wealth of unmarried women, unmarried men, and married couples in the super rich. First, women earn less income than men overall; yet women have higher saving rates than men that may equalize wealth holdings for high-wealth unmarried people. Women’s education levels, labor force participation, representation in leadership positions in organizations, and other measures of advancement have all increased in recent decades (Blau & Kahn, 2017; Schwartz, 2010). Despite these gains, however, there are still significant gender gaps in income across the income distribution (Blau & Kahn, 2017; Dinovitzer et al., 2009; Raley et al., 2006; Schwartz, 2010). Gender differences in income are particularly pronounced among those with the highest salaries, reflecting women’s underrepresentation in highly-compensated occupations (Dinovitzer et al., 2009; Neely, 2018; Rivera & Tilcsik, 2016). Moreover, there are more women than ever in top leadership positions today, but women are still underrepresented in these positions as well, and these differences intensify over the lifecycle (Warner et al., 2018). As a result, women’s income is sufficient to put only 1 in 20 households in the one percent by income (Yavorsky et al., 2019). Higher education — particularly having a professional degree — increases women’s incomes (England, 2010; Hout, 2012) and their odds of being in the one percent by income based on their own careers (Yavorsky et al., 2019). Importantly, however, income and net worth are correlated at only about .50 suggesting that saving from current income can overcome even sizable income gaps (Keister & Lee 2017; Keister, 2018), and high-income, high-wealth women spend less and save more from current income than men (Stanley, 2005). Women are more likely to save for known expenses, to shop at discount stores, and to otherwise find ways to be frugal than their male counterparts, even when they have sufficient income to meet their basic needs (Stanley, 2005). Together, these income and saving differences may equalize wealth ownership for wealthy unmarried women and men.

Second, high-wealth married couples tend to prioritize men’s careers; as a result, having two adults in a household does not necessarily double the household’s wealth holdings. The majority of couples now have two incomes, and educational and income homogamy have increased dramatically. Yet couples still prioritize men’s careers over women’s, at least partly because men have more leadership opportunities and are more likely than their wives to have opportunities to take jobs with very high incomes (Blau & Devaro, 2007; Cooke et al., 2009; Weeden et al., 2016). This tendency is particularly pronounced once a couple has children (England, 2011; England et al., 2016; Stone, 2007). When income gaps emerge in couples, women — including successful, ambitious women — may reduce their paid work (Sayer et al., 2009; Stone, 2007; Yavorsky et al., 2020). Public policies and gender norms reinforce that prioritizing men’s careers is ideal and that women are better at domestic work and childrearing and men are better at paid work (Thébaud, 2010; Thébaud & Halcomb, 2019). Related norms encourage women to use their household division of labor to signal social status: for high-income, high-wealth couples, a traditional division of labor may be seen as a luxury and a measure of prestige (Tichenor, 2005; Yavorsky et al., 2020). These processes work together to lead super-rich couples to opt for a traditional male breadwinner-female homemaker/caregiver arrangement much more frequently than other couples (Yavorsky et al., 2020). Accordingly, such patterns might lead to important differences in asset ownership that create and maintain wealth differences among unmarried women, unmarried men, and married couples.

Third, women take fewer investing risks than men, and women start businesses and invest in other business assets at lower rates than men. Importantly, men make the majority of investing decisions in married couples, which may lead high-wealth married couples to invest more like high-wealth unmarried men than unmarried women (Cantwell, 2014; Sherman, 2017) Women are starting businesses at unprecedented rates today, but they still start businesses at lower rates than men and persist in those new ventures for shorter stretches than their male counterparts. These patterns stem, in part, from the challenges women face in the entrepreneurship space. Specifically, women, compared to men, face greater barriers securing financial capital to start their businesses, and once they have started businesses, raising enough capital to grow a business past its nascent stages (Cantwell, 2014; Renzulli et al., 2000; Yang & Aldrich, 2014). The latter may limit the size of a woman’s business, the income it generates, and the assets that can be sold over time (National Women’s Business Council, 2012; Saurav et al., 2013; Warner, 2014). Moreover, because of women’s disproportionate family responsibilities (Yavorsky et al., 2015), many women start businesses to create greater work-family flexibility, in contrast to men who are more likely to start a business to advance their careers (Jennings & Brush, 2013). Such patterns may translate into different strategies for building the business and its associated assets. Entrepreneurship — and its resources and benefits — are often passed intergenerationally (Aldrich et al., 1998), suggesting that gender differences in one generation’s entrepreneurship may exacerbate gender differences in the next. Importantly, business start-up is a common path to the one percent (Edlund & Kopczuk, 2009; Keister, 2014; Keister & Lee, 2014; Yavorsky et al., 2019) and business assets are a critical part of the wealth portfolio of those in the one percent (Keister, 2014; Piketty, 2013; Saez, 2013). In married couples, men tend to have more influence over investing decisions than their wives (Chang, 2010; Sherman, 2017; Stanley, 2001 & 2005). It follows that there will be differences in investing strategies between unmarried women and unmarried men, but the wealth portfolios of unmarried men will resemble those of married couples more than those of unmarried women.

Fourth, women live longer than men. This simple demographic fact implies that women are more likely to be widows and to live longer than men as widows with ownership of formerly — and potentially high-value — marital assets. Life expectancy has increased dramatically since the start of the twentieth century for both genders, but the increase has been more pronounced for women than for men. In 2020, women’s life expectancy is 81 years, whereas men’s life expectancy is 76 years (Crimmins & Zhang, 2019; Duffin, 2020). This may seem like a small difference, but the disparity is large enough that women will spend notably more years as retirees and as widows. It follows that whereas men are more likely to have resources from current income and business start-up to save and build assets, women may potentially rely more heavily on transfers from their families of origin (inter- or intra-generational transfers) or their spouses who have preceded them in death. Consistent with this, previous research has assumed that women are more likely to inherit than to build their own wealth, due to a variety of barriers in the workplace, occupational segregation, and unequal divisions of labor in the home. This work finds, for example, that the percent of women in top wealth households followed an inverted U-shaped curve, peaking in the late 1960s because self-made wealth became more common in the late 1960s (Edlund & Kopczuk, 2009). This is also consistent with research that shows that women’s main route to the one percent is through their husband’s income — and most likely work-related financial benefits (Yavorsky et al., 2019).

3 Hypotheses

Together, these patterns suggest that women and men are likely to have different paths to top wealth positions, and that these paths will result in notably different wealth levels and wealth portfolios for those who reach the top. Several hypotheses follow. First, because men tend to have higher incomes than women and married couples prioritize men’s careers, unmarried men and married couples are likely to be able to save and accumulate more assets than unmarried women. However, because women have high savings rates and because a traditional division of labor limits the total income and savings of married couples, the differences in saving among unmarried women, unmarried men, and couples are likely to be modest. That is, among those in the top one percent of wealth owners, we expect that:

Hypothesis 1. There are modest differences in total net worth among unmarried women, unmarried men, and married couples with unmarried women owning slightly less net worth than either unmarried men or married couples.

There are also likely to be differences in the assets held by high-wealth women, men, and married couples. Business assets are a critical component of the assets of most wealthy families (Keister et al., 2021). Because men take more investing risk than women and invest more in other business assets (Yilmazer & Lyons, 2010), unmarried men are likely to have more of their assets in business forms of wealth. In addition, because men tend to make more financial decisions in married couples (Carman & Hung, 2017), married couples are likely to have risk preferences and investing strategies that resemble those of unmarried men more than they resemble those of unmarried women. Moreover, among recently widowed or divorced men and women (who we categorize as unmarried), women may be more disadvantaged in asset management after a marriage ends or their spouse passes away. Recent research suggests that individuals less involved in major financial decisions before a marriage ends or their spouse passes away may have less knowledge or skills to manage or maintain business assets acquired during the marriage (Xu, 2019). Because women are rarely the primary breadwinner in these super-rich couples (Yavorsky et al., 2019) and thus less likely to be responsible for asset management before divorce or widowhood, their business assets may take a greater hit after they lose a partner (and their financial know-how) than men in similar situations. As a result, the wealth portfolios of unmarried men are more likely to resemble those of married couples than unmarried women. That is, among those in the top one percent of wealth owners, it is likely that:

Hypothesis 2A: Unmarried women hold a lower percentage of their assets as business assets than married couples.

Hypothesis 2B: Unmarried men hold a similar percentage of their assets in business assets as married couples.

In addition, it is likely that super-rich women own a higher percentage of total assets as trust accounts than business assets. Trust accounts are financial instruments used to hold assets transferred from one party to another for the benefit of the recipient. The nature, function, and sociological meaning of trusts has been explored in detail elsewhere (see, for example, Harrington 2017). Trusts are commonly used to transfer assets from parents or grandparents to children and grandchildren. Trusts are also used to transfer joint marital assets from the couple to a surviving spouse in the event that one member of the couple dies. A slightly less common use of trusts is a domestic asset protection trust that transfers marital assets to one spouse — usually the wife — in the event of a divorce, an acknowledgement that women often fair worse than men financially in a divorce because couples prioritize men’s careers. In each case, the advantage of the trust is that financial assets are transferred according to the wishes of the owner and allow the recipient to avoid resolving questions about the ownership of the assets in court (probate).

A growing body of research shows that trusts have been central to both macro- and microeconomic processes. Their role in macro processes is evident in how they facilitated the financialization of the U.S. and global economies (Harrington, 2012 & 2017; Krippner, 2005 & 2011). Trusts are a financial instrument that originated in Medieval England (Harrington, 2012 & 2017) and that have survived, in part, because of their elasticity (Maitland, 1936) and their role in allowing individuals and corporations to create one form of wealth from a previous form (Harrington, 2017; Krippner, 2011). Trusts facilitated financialization by encouraging profit maximization and capital mobility (Krippner, 2005) and by allowing economic actors to move wealth across international borders with little friction (Beaverstock et al., 2013). At the micro (or family) level, trusts have played an important role in wealthy families by allowing them to protect their assets from financial crises and to grow even during times of economic contraction (Harrington, 2012 &2017). More important for our purposes, trusts have also enabled wealthy families to retain their assets across generations (Harrington, 2017) by allowing wealthy men to take care of their widows and daughters (who could not otherwise inherit property) after the men died (Francis, 1791; Harrington, 2017; Maitland, 1936).

These historic patterns combined with contemporary gendered patterns of work, income, investing, and life expectancy differentials suggest that there are likely to be gender differences in the ownership of trusts. Never married women and never married men provide a simple example of likely gender differences in the ownership of trusts. Because men tend to have higher incomes and to invest in high-risk assets and business ventures at higher rates than women, a higher proportion of unmarried women may have reached top wealth positions because they inherited wealth via trusts rather than accumulating wealth through their own businesses or careers. This is consistent with research that shows that women are more likely to be in top income positions as a result of their husband’s work-related characteristics than as a result of their own (Yavorsky et al., 2019).

Changes in marital status are also likely to affect the types of financial assets that men and women own. For instance, widows are likely to have received a significant portion of their assets from former marital property, and many of these assets are likely to be transferred in trust accounts particularly for the very wealthy. Similarly, divorced women are also more likely than divorced men to have received former marital assets in the form of trusts. Moreover, trusts are managed by trustees who are legally required to act in the best interest of the recipient, and widows or divorced women often appoint financial professionals as trustees and outsource the asset management to them. Consequently, holding trusts can be an effective way to preserve wealth for many widows or divorced women who might not have developed asset management skills during marriage. For this reason, it is also likely that widows or divorced women liquidate inherited assets which were not already intended to become trusts at the death of a spouse or at divorce and convert the proceeds to additional trusts. By contrast, marital status changes are less likely to affect men’s asset portfolios given that men are likely to continue active participation in asset management following the death of a spouse or a divorce and, thus, less likely to keep funds in trusts. It follows that, for those in the top one percent of wealth owners:

Hypothesis 3A: Unmarried women hold a higher percentage of their assets in trust funds than married couples.

Hypothesis 3B: Unmarried men hold a similar percentage of their assets in trust funds as married couples.

4 Methods

5 Results

Most important for our purposes, Table 1 shows that, consistent with Hypothesis 1, there are only modest differences in the mean and median net worth for unmarried women, unmarried men, and married couples in the one percent. Unmarried women have the lower mean ($17.14 million) net worth compared to unmarried men ($17.41 million) and married couples ($19.72 million). However, the difference between the mean for unmarried women and unmarried men in the one percent is not statistically significant. Moreover, the median net worth for unmarried women and unmarried men in the one percent is nearly identical ($11.47 million). These descriptive statistics indicate that married couples in the one percent have slightly more net worth than unmarried women and men, a pattern that is evident in both the mean and median net worth values but that is only modestly significant. Consistent with previous research on income differences in the one percent (Yavorsky et al., 2019; Yavorsky et al., 2020), Table 1 shows that unmarried men have higher mean and median income than unmarried women, slightly lower median income than married couples, and equivalent mean income to married couples. These patterns are consistent with unmarried men in the one percent having high variance in their incomes (Table 1) and being the dominant breadwinner in households in the one percent (Yavorsky et al., 2019; Yavorsky et al., 2020). Indeed, the standard deviation associated with both net worth and income for those in the one percent is higher for unmarried men than for either couples or unmarried women, which suggests that unmarried men’s net worth is more dispersed over a wider set of values than the other groups.

Our descriptive results also provide initial support for our remaining hypotheses. Table 2 shows how the percent of gross assets held as financial assets and non-financial assets varies by gender, marital status, and position in the wealth distribution (top one percent versus other households). Those in the one percent of wealth holders have more financial assets, including business assets and trusts, than those in the remaining 99%. More relevant to our study, this table shows that there are notable differences in the wealth portfolios of unmarried women, unmarried men, and married couples. In particular, Table 2 shows that, consistent with Hypothesis 2A, among households in the top one percent, unmarried women (21.0%) hold a lower percentage of their net worth as business assets compared to married couples (37.8%), but unmarried men (35.3%) and married couples have relatively similar portions of the assets in business-related wealth, consistent with Hypothesis 2B. Table 2 also shows that among households in the top one percent, unmarried women (15.6%) hold a higher percentage of their net worth in trust funds, compared with married couples (3.5%; Hypothesis 3A), whereas unmarried men (4.3%) and married couples hold relatively similar portions of their portfolios in trust funds. Table 2 shows that there are other differences in the wealth portfolios of women, men, and married couples: for instance, unmarried women (14.6%) in the one percent have more of their assets held as stocks than unmarried men (13.2%) or married couples (10.3%). However, this difference is marginal given the more extreme differences across these groups in the ownership of trust accounts and business assets. Again, although these descriptive estimates provide initial support for our expectations, the values shown in Table 2 are preliminary and do not control for other behaviors and processes that are associated with wealth ownership and the allocation of assets across financial instruments.

Additional descriptive evidence included in Table 2 underscores differences in saving and investment risk for unmarried women, unmarried men, and married couples. Households in the one percent save for known expenses and take more investment risk than households in the remaining 99%. More important for our purposes, these descriptive statistics also show that there are meaningful differences in both measures among households in the one percent. In particular, among those in the one percent, unmarried women (38.7%) are more likely than unmarried men (30.1%) to save for known expenses and have more similar rates to married couples (43.6%). By contrast, unmarried men in the one percent (37.1%) are more willing than unmarried women in the one percent (22.4%) to take investment risk and have more similar rates to married couples (41.8%). In both cases (saving for known expenses and willingness to take investment risk), married couples have higher rates than unmarried men and women. Our data do not allow us to explore in greater depth the meaning of the high rates of both measures (saving and willingness to take investment risk) in married couples relative to unmarried persons, although this pattern suggests a financial advantage of marriage. Future research might usefully disentangle this pattern. Additionally, the fact that unmarried men are more similar to married couples in their likelihood to take investment risks is consistent with our suggestion that married couples are likely to have risk preferences and investing strategies that resemble those of unmarried men more than they resemble those of unmarried women.

Our multivariate models build on the descriptive statistics and provide additional evidence that there are differences among unmarried women, unmarried men, and married couples in total net worth. Table 3 includes two ordinary least squares models using total household net worth as the dependent variable. Total household net worth is divided by $1,000,000 for clarity. Model 1 includes only our measures of joint gender-marriage-wealth status (i.e., it omits all control variables). Model 2 introduces controls for age, education, race, and employment status. Both models show that there are modest differences in total net worth when unmarried women, unmarried men, and married couples are compared. That is, the coefficient for top one percent unmarried women shows that the net worth of top one percent unmarried women is significantly less than the net worth of top one percent married couples (reference). The coefficient for top one percent unmarried men shows that the net worth of top one percent unmarried men is less than that of top one percent married couples, but the difference is not statistically significant. Moreover, the coefficient for unmarried men in the one percent is greater than the coefficient for unmarried women in the one percent (i.e., it is less negative), indicating that among those in the top one percent, the net worth of unmarried men and married couples is more similar than the net worth of unmarried women and married couples.

We also find significant gender differences in the ownership of business assets and trusts in the super rich. Table 4 includes six models using the percent of total assets held as business assets (models 1–3) and the percent of total assets held as trusts (models 4–6). Models 1 and 4 include only an indicator that the responding household is in the top one percent by net worth (reference is all other households) and control variables for age, education, race/ethnicity, and labor force participation. Models 2 and 5 include measures that the responding household is an unmarried woman or an unmarried man in the top one percent (reference is married couples), and models 3 and 6 include interactions between being in the top one percent by wealth and being an unmarried woman or an unmarried man.

Our findings show that there are substantial differences in the percentage of assets held as business assets for those in the one percent (Model 3 of Table 4). The interaction term between top 1% wealth and unmarried women is -9.505, meaning that unmarried women in the one percent hold a lower percentage of their assets as business assets than married couples in the one percent (Hypothesis 2A). This finding is consistent with literature that finds men invest more in business assets than women (Yilmazer & Lyons, 2010). Previous research did not focus exclusively on the super rich, but it is logical that these patterns would be heightened in the top one percent given that business assets are much more common in the portfolios of those in the one percent (Table 2) and that men tend to make more financial decisions in married couples (Carman & Hung, 2017). In supplementary analyses, we break down our unmarried groups further into widowed, divorced/separated, and never married women and men (see Appendix Table A). The findings in these additional analyses help explain the main patterns displayed in Table 4. For example, we find that separate/divorced and widowed appear to be driving the finding that unmarried women hold a lower percentage of business assets. Among recently widowed or divorced men and women (respondents who are included as unmarried in our main analyses), women may face greater difficulties in managing their assets when a marriage ends or their spouse passes away and/or their net worth may decline because they were less involved in these financial decisions when they were married (Xu, 2019). In the case of divorce, it may be the case that the dissolution of marriage spurs the separation of business assets between the former spouses, reducing the net worth of both parties.

Similarly, because men tend to be the primary breadwinners in super-rich couples and tend to play a more significant role in financial management in these couples, we anticipated that unmarried men in the one percent would hold a similar percentage of their assets as business assets as married couples (hypothesis 2B). As anticipated, the interaction term between top 1% wealth and unmarried men (-4.684) is not statistically significant, supporting hypothesis 2B. Note that when we examine the more discrete categories of unmarried men, we find an exception with separated/divorced men. As shown in Appendix Table A, separation or divorce appears to hurt men’s business assets too (see the negative and significant interaction coefficient for top 1% wealth and separated/divorced men), like it does for women.

Our results also show that there are notable differences in the percentage of assets held as trusts for those in the one percent (Model 6 of Table 4). As the interaction term between top 1% wealth and unmarried women (10.296) shows, unmarried women in the one percent hold a higher percentage of their assets as trusts than married couples in the one percent (Hypothesis 3A). Some of this effect likely reflects changes in financial assets that occur following a change in marital status. For example, in our additional analyses, we find that the interaction between top one percent wealth and divorced/separated women is significant and positive. Thus, it appears that divorced/separated women may liquidate inherited assets and move the funds to trust accounts. Finally, results also provide evidence consistent with our expectation that, among those in the one percent, unmarried men hold a similar percentage of their assets in trust funds as married couples (Hypothesis 3B). The interaction between top one wealth and unmarried men is positive but is not statistically significant. Again, because men’s financial decision making and investment strategies are likely to dominate those of married couples in the super rich, we anticipated that unmarried men and married couples in the one percent would have similar percentages of their assets held as trusts.

Figure 2 uses predicted percentages for business assets and trusts to illustrate differences in the portfolios of unmarried women, unmarried men, and married couples in the one percent. Using Models 3 and 6 from Table 4, we calculated the predicted percentages for every observation in the sample by manipulating the values corresponding to the three super-rich groups (unmarried women, unmarried men, married couples in the one percent) while retaining original values for other covariates. The calculated predicted percentages were then averaged across all observations. We also added 95% confidence intervals for the predicted percentages to Figure 2. The figure illustrates that the significant differences in the portfolios of the three super-rich groups are consistent even after controlling for relevant covariates. Unmarried women in the one percent have a lower percentage of business assets but a higher percentage of trusts than married couples in the one percent. By contrast, unmarried men in the one percent have a similar percentage of those assets as married couples in the one percent.

Exploratory analyses showed that the patterns we describe here are largely consistent over time. That is, there have been few changes in these patterns in recent decades. We do not include interactions between year and our key test variables because the results of these exploratory models indicated that our findings have been relatively unchanged in the years covered by the SCF. Future research might usefully explore these patterns in more detail and might examine why there has been little change. Indeed, the lack of change is important for understanding women’s position in the stratification system and the gender revolution; future research might extend these findings to address such issues.

6 Discussion and Conclusion

This article studied the super rich by exploring gender differences in the wealth levels and the portfolio composition of the top one percent of U.S. wealth holders. We proposed that gender and gendered family dynamics are likely to lead to differences in net worth and asset ownership when unmarried women, unmarried men, and married couples in the one percent are compared. By looking at these three groups, we were able to compare respondents by gender and treated married couples as joint owners of shared marital property, a challenge that has stalled previous research on gender and wealth. We focused on three outcomes: total net worth, the percent of assets held as business assets, and the percent of assets held in trusts. Together these measures allowed us to provide a glimpse into the overall wealth of households and to explore differences in the prevalence of two key assets owned by the super rich. Our results were consistent with our expectations regarding overall wealth owned by the one percent: we found that there were modest differences in net worth among unmarried women, unmarried men, and married couples in the one percent; we also found that unmarried women in the one percent own slightly less net worth than either unmarried men or married couples in the one percent. We also found empirical support for our expectations regarding the prevalence of business assets and trusts in the wealth portfolios of the super rich. That is, we found that among those in the one percent, unmarried women hold a lower percentage of their total assets in business assets compared to married couples; we also found that unmarried men in the one percent hold a similar percent of their assets in business assets as married couples in the one percent. Finally, our results provided evidence for our expectation that among those in the one percent, unmarried women hold a higher percentage of their assets in trust funds than married couples, while unmarried men hold a similar percentage of their assets in trusts.

These findings provide a glimpse into the processes that give people access to the highest financial positions in the United States. It has been clear from other research that the super rich have many more business assets than other households (Benton et al., 2017; Keister, 2014; Nau, 2013); anecdotal evidence and evidence from lists of the super rich (e.g., the Forbes 400) also suggest that business assets are more common among those at the top of the wealth distribution (Edlund & Kopczuk, 2009; Kroll, 2018; Freeland, 2012). However, previous research has not isolated the super rich and explicitly examined the proportion of their assets that are held as business assets. Our estimates fill this gap and, in doing so, show clearly that business ownership and investment in business assets is a key pathway to top wealth positions. Moreover, we find that the ownership of business assets is a more important correlate of membership in the one percent for unmarried men and married couples than for unmarried women. Trust funds are another important financial instrument for the very wealthy (Harrington, 2017; Khan, 2011; Lerner et al., 1996). Our work builds on this research and suggests there are important gender differences in the way trusts are held by the one percent. Unfortunately, we cannot say with any additional certainty that the unmarried women in our sample received their trusts from their deceased husbands, following a divorce, or from their wealthy parents or grandparents. Our data do not include additional information about the source of the trust fund; our data are also cross sectional, making it impossible to study whether the same respondent had no trust accounts in one year and, for example, added a trust fund to their assets following the death of a spouse. Finally, our research cannot address whether historic patterns in the use of trusts underlies the relationships we find here. Future research could explore these issues.

Our findings also provide a unique approach to understanding the status of women at the very top and may contribute to research that studies whether women’s position has improved over time at various economic distributions. The literature on gender and work and financial behavior has been slightly controversial. Women’s positions have, indeed, improved in some important respects: education levels, incomes, entrepreneurship, and representation in top leadership positions have all increased for women in recent decades. Yet pronounced gender disparities remain (Cotter et al., 2001; England et al., 2020; Rivera & Tilcsik, 2016; Yavorsky et al., 2019). Among the one percent, it has become clear that women’s incomes are rarely sufficient to push a household into the one percent by income (Yavorsky et al., 2019) and that super-rich couples prioritize men’s careers and contributions to the household by having a traditional division of labor (Yang & Aldrich, 2014; Yavorsky et al., 2020). Our work builds on this scholarship by showing that women in the one percent of the wealth distribution are more likely to be in these super-rich positions as a result of someone else’s work- or business- related contributions than their own. Our findings also build on the growing body of sociological research on trust accounts (Harrington, 2017) to show that super-rich unmarried women are more likely to have trusts than unmarried men or married couples, and super-rich unmarried women are less likely to have business assets (assets that are likely to have been self-made) than unmarried men or married couples. Of course, class advantage facilitates business start-up and the acquisition of business assets, and not all trusts reflect passive wealth transfers. However, our findings are suggestive of a gender pattern that is consistent with the assumptions of prior research (Edlund & Kopczuk, 2009; Rosenfeld, 1998) and that suggests that women may be dependent on others for access to the super rich.

These gendered patterns are very likely reinforced by rising inequality trends — where the super rich have continued to move further and further away in terms of financial resources from the average American or those in other rich positions (in the 80^th–89^th percentile or 90^th to 99^th percentile) (Balestra, 2018; Keister & Lee, 2017; Killewald et al., 2017; Bhutta et al., 2020). Whereas women have made progress entering professional jobs, like medicine and law, they still remain rare among the highest paying specialties within these fields or top entrepreneur positions (Warner et al., 2018; England et al., 2020). As the super rich pull away from the bottom 99%, driven by (predominately white) men’s financial resources, the progress other historically marginalized groups (women and people of color) have made cannot compete or keep pace with the amount of resources it takes to be in top 1%. Given that economic, political and social power typically accompanies top 1% positions, particularly for breadwinners, it is critical to better understand this elite group and how people access it.

Our research highlights the different wealth portfolios of the super rich, based on different marital statuses and gender. To be clear, we are not arguing that unmarried (or married) women in the super rich are disenfranchised but rather highlighting how larger patterns of gender and class influence the make-up of the top one percent, with broader implications for who controls the majority of wealth in the U.S and in what form. We also acknowledge that there are important ways in which this work could be extended. For example, future research could usefully explore whether there are differences across households in the top 20% of wealth holders, comparing those who are in the top one percent to other high-wealth households who are not quite wealthy enough to be in the top one percent. It is likely that gender patterns of wealth ownership differ even within the top 20 percent in ways that could usefully inform understanding of inequality and gender patterns of wealth ownership. For instance, wealthy couples who are just below the one percent are typically younger than those at the very top. Exploring how their wealth ownership differs from that of couples in the one percent could be suggestive of age and cohort differences and could inform understanding of mobility over the life course across the wealth distribution.

Future research might also explore how gender interacts with other variables that are important to understanding inequality, including labor market status and education to produce patterns of wealth ownership. For example, educational attainment has increased for women (England et al., 2020), and women are now more likely than ever before to have leadership positions, including in large corporations (Warner et al., 2018). It follows that patterns of wealth might vary for women based on their human capital. Women’s changing position relative to their spouses is another key dimension on which gender patterns of wealth ownership might vary. In this spirit, future research might also explore whether educational, age, and labor force homogamy intervene between gender and wealth ownership. Each of these ideas could be accomplished with the SCF data we use in this paper. Beyond these ideas, future research might also explore the role of social capital — or social relations — in mediating the gender-wealth relationship. In other words, there is some potential that the social capital that women and men develop during college and graduate school and in the workplace differ. To the extent to which wealth accumulation (e.g., business ownership) is more likely and successful with the right social capital, there may be differences in our findings. Unfortunately, the SCF does not include data on social capital, but today’s network data collection and analysis methods would make answering questions of this sort feasible.

It is also important to mention that the super-rich hold enough political power to influence measures of wealth taxation. Indeed, there are tax incentives for elite families to use trusts to reduce wealth when both spouses are alive and after the death of one spouse. These measures could potentially affect the distribution of wealth within the household in ways that are reflected in the use of trust accounts. Although the SCF does not contain sufficient data to explore this possibility, future research might use other data sources to explore the use of various wealth reducing trust accounts in shaping elite wealth. An even more ambitious research agenda might also explore the extent to which elites manage to influence public policies regarding trusts and related issues such as estate tax levels in ways that directly benefit their personal net worth and gendered patterns of wealth ownership.