Financial Solidarity or Autonomy? How Gendered Wealth and Income Inequalities Influence Couples’ Money Management

It is well established that women have lower income and wealth levels than men. These inequalities are most pronounced within heterosexual couples and grow once partners get married and have children. Nevertheless, equality in control‐ ling money within couples is highly valued and might ameliorate women’s disadvantages in income and wealth ownership. Previous research has focused on explaining gender wealth inequalities at the household level; less is known about the pos‐ sible consequences of these inequalities on how couples manage their money. In this article, we investigate how income and wealth inequalities among couples are associated with joint or independent money management. In theoretical terms, we perceive money management systems as representing two different norms of reciprocity within couples for buffering income and wealth inequalities between partners, depending on the transferability of resources and their institutional reg‐ ulation. We apply pooled logistic regression models to data from the German Socio‐Economic Panel Study. Our findings confirm that income and wealth are relevant but have opposite associations with couples’ money management strategies. While couples with unequal income constellations tend to pool their money, couples with unequal wealth constellations manage their money independently. Accordingly, couples seem to use labour income to buffer gender inequalities by shar‐ ing resources, thereby following the norm of partnership solidarity. In contrast, gender wealth inequalities are reproduced by keeping resources separate, thus representing the norm of financial autonomy.


Introduction
Labour income and wealth are important dimensions of social stratification in contemporary societies (Hällsten & Thaning, 2022;Killewald et al., 2017;Pfeffer & Waitkus, 2021;Skopek et al., 2014). This particularly holds true for gender inequalities. In most Western industrialised countries, women not only earn less income than men, they also possess less wealth: Compared to other OECD countries, the gender wage gap in Germany is especially high (OECD, 2018). As a consequence, partnered women contribute significantly less to the household income than their male partners (Dieckhoff et al., 2020;Krause, 2008). Wealth disadvantages for women in Germany are even larger, and again, they are most pronounced for partnered women Schneebaum et al., 2017;Sierminska et al., 2010). Thus, couples' financial arrangements are central to understand the (re-)production of gender inequalities. However, both women and men also benefit from partnerships in terms of wealth accumulation. Partnered women and men not only have more long-term financial security due to overall higher wealth levels but are also able to reduce gender inequalities in income and wealth by sharing access to resources or joint investments (Frémeaux & Leturcq, 2022;Lersch, 2017b;Nutz, 2022).
Empirical research on money management in couples has centrally focused on differentiating between the ownership of and access to money. Individual access to money in couples depends on how the partners manage and control their money (Evertsson & Nyman, 2021;Pahl, 1989). In Germany, about two-thirds of couples pool their money and jointly manage it, followed by about onefourth of couples who engage in independent management; patterns in which one partner is solely responsible are less common today (Çineli, 2022;Lott, 2009Lott, , 2017. In this article, we ask why couples pool their money or manage it independently. Previous research shows that partners' income constellations are an important factor in understanding couples' money management (Çineli, 2022;Lott, 2009;Ludwig-Mayerhofer, 2006). This article expands these findings by examining whether wealth inequalities between partners, in addition to income inequalities, are also associated with their money management. Even though couples generally do not pool their wealth in terms of ownership (Lersch, 2017a;Nutz, 2022), equality in control and management of money within couples is also highly valued in terms of wealth (Evertsson & Nyman, 2021;Joseph & Rowlingson, 2012;Tisch & Lersch, 2021).
Conceptually, we differentiate between labour income and wealth as distinct monetary resources. While labour income is mainly used to cover couples' daily living expenses and is easily transferable between partners, wealth provides an important basis for financial security and larger expenditures, that is, long-term investments (Hällsten & Pfeffer, 2017;Rodems & Pfeffer, 2020). Wealth thus serves as a buffer against couples' life risks in the long run. Compared to income, wealth ownership is less easily transferable between partners and more strongly regulated by legislation. We further divide couples' wealth into the components of home ownership and non-housing wealth, since home ownership is a key determinant of wealth inequality in general (Pfeffer & Waitkus, 2021;Skopek et al., 2012Skopek et al., , 2014. Moreover, gender inequalities differ significantly between these two wealth dimensions (Joseph & Rowlingson, 2012;Kapelle, 2022;Nutz & Gritti, 2021).
In the following, we present the state of research on gender inequalities in labour income and wealth as well as on money management within couples. Thereafter, we develop a theoretical framework based on exchange theory on how couples' income and wealth inequalities might be associated with strategies of pooling and independently managing money. We explore these associations empirically with data from the German Socio-Economic Panel (SOEP) by estimating pooled logistic regression models for a sample of heterosexual couples.

Gender Inequalities in Income and Wealth and Couples' Money Management
The gender wage gap, that is, the fact that women earn less than men, is a stable feature of Western industrialised countries (OECD, 2018). Compared to other OECD countries, Germany has a particularly high gender wage gap, with women having earned about 18% to 22% less than men since the mid-1990s (OECD, 2018;Statistisches Bundesamt, 2022). The few empirical studies examining wage differentials within households have indicated that partnered women in particular earn less than their male partners (Dieckhoff et al., 2020;Krause, 2008). Previous research has highlighted that gendered employment careers and care responsibilities are especially important for explaining gender inequalities in wages and income, particularly within couples (e.g., England et al., 2020;Minkus & Busch-Heizmann, 2020).
Wealth is even more unequally distributed between men and women than labour income. In a European comparison, the gender wealth gap in Germany is rather large and mainly driven by gender disparities in gross wealth (Schneebaum et al., 2017), with women possessing about 28% less wealth than men (Grabka & Westermeier, 2014). Again, gender wealth inequalities are most pronounced within couples Sierminska et al., 2010) and are mainly due to women's substantially lower accumulation of financial wealth (Kapelle, 2022). While financial assets are more often held separately (Nutz & Lersch, 2021), housing wealth is predominantly owned jointly by both partners and represents the largest investment for most couples (Joseph & Rowlingson, 2012;Kapelle, 2022;Nutz & Gritti, 2021).
Most studies on gender wealth inequalities, both in general and within couples, have investigated how these disparities come about and have identified four important factors: First, gender wealth inequalities are driven by men's higher labour incomes Sierminska et al., 2010). Second, the gender wealth gap seems to be highest within married couples (Killewald et al., 2017;Sierminska et al., 2010), particularly for older cohorts and for non-real-estate wealth (Lersch, 2017b). Third, the birth of a child seems to have a negative impact on women's wealth development, while men's wealth is hardly affected by it (Lersch et al., 2017). Fourth, the matrimonial property regime seems to matter . While married couples with a community property regime accumulate less wealth than married couples with a separate property regime, the gender wealth gap is smaller among couples with a community property regime (for data from France see Frémeaux & Leturcq, 2022). However, married couples only very rarely opt out of the default property regime of accrued gains in Germany .
While most quantitative studies have analysed gender inequalities in the ownership of wealth, qualitative studies have revealed that the possession of and access to wealth within a couple might in fact not only depend on ownership but on the perceptions of who owns their wealth (Evertsson & Nyman, 2014;Joseph & Rowlingson, 2012). This finding is supported by the seminal study by Pahl (1989) for the United Kingdom, which showed that money flowing into the household was not automatically regarded as the total income of the family; rather, access to money depended on how partners managed and controlled their money. She identified different money management systems, depending on whether the man or the woman was solely responsible, whether both partners pooled their money or whether they managed it independently (Pahl, 1995(Pahl, , 2005. Comparative studies on couples' money management indicated that today, the pooling system is generally adopted by a majority of couples; independent management predominates in very few European countries, such as Finland or Portugal, while sole money management, mostly by men, is more prevalent in non-European countries (Çineli, 2022;Yodanis & Lauer, 2007). Germany represents a rather typical case among European countries, with 64% of couples completely pooling their money, 24% managing all or most of their money independently, and only 12% assigning sole responsibility to either the man or the woman (with data for 2005 see Lott, 2009, p. 339; see also Holst & Schupp, 2005;Lott, 2017;Ludwig-Mayerhofer, 2006).
The partners' relative labour income has been identified as an important factor influencing couples' money management. Comparative studies have indicated that couples with income homogamy either adopt the pooling system or manage money independently rather than engaging in sole management (Çineli, 2022;Yodanis & Lauer, 2007). In Germany, pooling is more common if the partners' income constellations are very unequal. In contrast, if women's income share increases, independent management becomes more likely (Hu, 2021). Empirical studies examining changes in money management within couples are rare, but their results support the evidence from analyses of between-couple differences (Lott, 2017).
Previous literature on money management mainly refers to the management of money derived from labour income. The importance of partners' wealth constellations for couples' money management has not yet been studied. Results from qualitative studies have pointed towards a relevant association since couples with unequal constellations refer to each other's wealth as a buffer when explaining their money management practices (Evertsson & Nyman, 2014;Joseph & Rowlingson, 2012). Therefore, we investigate how wealth inequalities within couples, in addition to income inequalities, are associated with how couples manage their money in Germany. In the following, we develop a theoretical framework based on exchange theory and then derive hypotheses in this regard.

Norms of Reciprocity and Money Management Systems in Couples
We start from the premise that today, most couples no longer opt for the traditional sole management system but have to decide on whether to adopt a complete pooling or an independent management system. To explain couples' money management, previous research has often applied resource theory (Blood & Wolfe, 1960), which assumes that the more resources a person has in terms of education, income, or occupational status, the more power he or she will have in the relationship. Applied to couples' money management, this implies that the partner with the greater quantity of resources would be solely responsible for managing the money of the household (cf. Çineli, 2022;Lott, 2009;Yodanis & Lauer, 2007). However, resource theory is less suitable for understanding independent management systems (Hu, 2021;Lott, 2017), since it cannot explain why the partner with the lesser quantity of resources retains the power to manage money on his or her own within the couple.
We therefore apply a broader exchange perspective and perceive couple relationships as exchange relations (for an overview see Lott, 2012, pp. 42-56; see also Safilios-Rothschild, 1976). Within the framework of exchange theory, spouses exchange different types of desirable goods, which include both economic and non-economic resources, such as financial support or material goods, but also love, esteem, and respect (Safilios-Rothschild, 1976). Partners exchange these valued resources according to established rules of exchange, mostly according to the norm of reciprocity, since the long-term balance of exchange is more important than the gains of one partner (Lott, 2012, p. 44). At the same time, the availability of alternatives to each partner that can provide these valued resources is also important (Safilios-Rothschild, 1976). Resources such as individual income and wealth, but also access to money and the division of labour within the couple, are therefore part of the exchange process and establish a couple-specific dependence relation. Based on exchange theory, we perceive the pooling and independent money management systems as representing varying notions of reciprocity in couples' exchange processes (cf. Pepin, 2019).
The pooling system involves the complete sharing of the couple's money and its management (Hu, 2021;Pahl, 1995;Vogler et al., 2006). Both partners have equal access to all money that enters the household, and both spend from a common pool, such as a joint bank account. "Couples adopting this system often explain that 'it is not my money or his/her money-but our money' " (Pahl, 1995, p. 366). The norm of reciprocity associated with the pooling system is called "partnership solidarity," which prioritises the shared goals of both partners through the sharing of resources. This norm of solidarity treats the couple as a unit, where both gains and losses are shared equally between partners, especially in unequal constellations with regard to economic resources (Hu, 2021;Pepin, 2019). Having equal access to all the couple's resources offers the potential for equal sharing and therefore might serve as a buffer against life risks for both partners. At the same time, this system does not necessarily indicate effective equal control of money, since it often conceals the fact that the partner who is earning less money, mostly the woman, often has (self-)restricted access to the money from the joint pool (Vogler et al., 2006). The independent management system "is defined by both partners having their own source of income and neither having access to all the household funds" (Pahl, 1995, p. 366). Accordingly, couples managing their money independently base their exchange on the reciprocal norm of "financial autonomy for both partners." They perceive their partnership as an association of two autonomous individuals who retain ownership of their respective money, and they are oriented towards maintaining economic independence for both partners (Bennett, 2013;Hu, 2021;Pepin, 2019;Vogler et al., 2006). However, this is not at odds with couples perceiving themselves as a unit and pursuing collective goals (Nyman, 2003;Pahl, 2005). Couples embracing the reciprocal norm of autonomy often strive for (financial) equality. However, independent money management does not necessarily guarantee equality but provides financial autonomy to those individuals who earn enough to make a living on their own (Evertsson & Nyman, 2014). Therefore, if couples with unequal resource constellations manage their money independently, they often negotiate their finances and employ informal compensation payments (Evertsson & Nyman, 2014;Joseph & Rowlingson, 2012;Nyman, 2003). Consequently, life risks remain more individualised and protection against life risks is more strongly related to partners' individual resources than in the pooling system (Hu, 2021;Vogler et al., 2006).
In sum, we perceive these two different money management systems as representing specific norms of reciprocity that are associated with different ways of buffering inequalities between the partners (Evertsson & Nyman, 2014;Nyman, 2003;Tisch & Lersch, 2021). If labour income and wealth constitute valued economic resources in the exchange process, couples' use of either the complete pooling or the independent money management systems should be linked to the corresponding norm of reciprocity. However, since income and wealth differ regarding their transferability between partners and the legal regulations governing them, we assume that the specific norm of reciprocity adopted should differ for income and wealth inequalities within the couple.
Money from labour income is the most liquid financial resource; it is mainly used for living expenses in everyday life and can thus be regarded as a type of monetary flow (Killewald et al., 2017;Spilerman, 2000). It can be easily transferred between partners and directly used to make up for financial inequalities within the couple, i.e., by establishing a joint bank account from which both partners can withdraw money. We assume that couples adopt a pooling system and thus follow the norm of partnership solidarity if their income constellation is unequal, that is, if one partner earns substantially more than the other.
This should hold true for male breadwinning couples, which are based on the notion of separate spheres and according to which the male partner earns the income and the female partner takes care of the household and children (Yodanis & Lauer, 2007). Since the female homemaker bears the burden of higher economic risks, the reciprocal norm of solidarity between the partners ensures redistribution of income via the pooling system. This norm of partnership solidarity is reinforced by the normative ideal of the modified male breadwinner model that underlies the conservative German welfare state, for example, through the joint taxation of married couples with very unequal incomes (Hipp & Leuze, 2015), derived social security entitlements for the partner with no or only a very low income (Minkus & Busch-Heizmann, 2020), or the community of accrued gains as the default matrimonial property regime . This should likewise hold true for female breadwinning couples. In these non-traditional constellations, both partners seek to comply with traditional gender roles too, since women still carry the burden of housework and childcare despite earning more than their partners (Brines, 1994). As a consequence, money from women's income should have less value in claiming financial control (Zelizer, 1989) and both partners may consider it legitimate for the man to exercise control over financial decisions (Tichenor, 2005). Therefore, in this income constellation, the reciprocal norm of solidarity between partners should also result in pooled management but mainly due to compliance with traditional gender roles. The pooling of money therefore should be more prevalent among couples with unequal income constellations, regardless of whether the male or the female partner earns more.
In contrast, couples with income homogamy do not have to rely on the reciprocal norm of solidarity to the same extent. Due to the dual-earner model, both partners share similar economic risks and both partners retain the economic potential to make a living on their own to a certain degree. This likely strengthens the norm of financial autonomy for both partners, which increases the likelihood of independent money management (Çineli, 2022;Lott, 2009). We therefore expect an unequal distribution of labour income between partners to increase the likelihood of a pooled management system, while a rather equal distribution of labour income should make the independent management system more likely (H1).
Wealth inequalities within couples should engender different associations with the type of money management. Unlike income, wealth represents a stock of financial resources (Killewald et al., 2017;Spilerman, 2000), since it constitutes an important basis for financial security in the long run and thus can reduce life risks in the long term (Hällsten & Pfeffer, 2017;Rodems & Pfeffer, 2020). Moreover, legal regulations and the design of contracts required for the ownership of wealth impose limits on the transferability between partners. This is most obvious in the case of real estate assets, including homeownership, which have the lowest liquidity and, if owned only by one partner, also the lowest level of transferability between partners (Hällsten & Pfeffer, 2017). But also financial assets, such as investment shares, are mostly owned by one person and only become available if sold on the financial market (Nutz & Lersch, 2021). They offer slightly more liquidity in the short run than real estate assets but are still less easily transferable between partners than labour income.
Moreover, there are no clear institutional incentives to pool wealth throughout the partnership, a feature that distinguishes wealth from income. Even in the default matrimonial property regime, the community of accrued gains, personal wealth remains the individual property of the two partners . In addition, wealth inequalities within couples are not visible in everyday routines but only become relevant in certain situations, for instance when couples separate (Boertien & Lersch, 2021;Kapelle, 2022). Persistent wealth inequalities within the couple often emerge from outside of the couple context-i.e., they are rooted in wealth differences from before union formation and result either from intergenerational inheritances and donations or from previous relationships or marriages (Fagereng et al., 2022;Joseph & Rowlingson, 2012). Therefore, the wealthier partner has a strong incentive to secure his or her individual long-term financial security, especially if one partner has higher debts. This does not preclude the wealthier partner from using their assets as a buffer for the less wealthy partner. Nevertheless, support and temporary redistribution are typically achieved by informal compensation payments rather than by changes in wealth ownership (Evertsson & Nyman, 2014;Joseph & Rowlingson, 2012). In couples with marked wealth inequalities, we thus assume that the norm of financial autonomy will prevail and make independent money management more likely.
In contrast, equal wealth distribution between partners mainly results from joint investments in assets (Nutz, 2022). Joint homeownership in particular is a common strategy of male breadwinning couples to compensate for unequal life risks in the long run that result from the division of paid and unpaid labour (Joseph & Rowlingson, 2012). This implies that both partners do not necessarily make the same financial contribution to the joint investment, but that ownership of the asset is shared equally. This might strengthen the norm of partnership solidarity through the sharing of resources. Therefore, we expect an equal distribution of wealth between partners to be associated with a higher likelihood of a pooled management system, while an unequal distribution of wealth should increase the likelihood of an independent man-agement system (H2). Overall, the lower the transferability between partners, the stronger the association between the gender wealth ratio and couples' money management should be. We thus distinguish between owner-occupied housing wealth, which should have a low degree of transferability compared to other financial resources, and non-housing wealth, which should be more easily accessible for spending purposes.

Data, Measurements, and Methods
For our empirical analysis, we used data from the SOEP (v36; see Goebel et al., 2019). We relied on four survey years containing information on wealth (2002,2007,2012,2017) and merged them with information on money management from subsequent waves (2004,2008,2013,2018). Our unit of analysis was the couple. We limited our sample to heterosexual couples living in one household, with both partners aged between 18-64 years. We selected respondents who were both born in Germany and had German citizenship to ensure a comparable institutional frame for income and wealth accumulation. We relied on the imputed personal labour income, household income, and personal wealth data provided by the SOEP survey team (Grabka & Westermeier, 2015). Personal labour income was set to 0 for non-employed respondents with missing values. We addressed item non-response affecting other relevant analytical variables through multiple imputation by chained equations using Stata's mi procedure (version 16); to do so, we combined estimation results from five imputed data sets using Rubin's rule (Rubin, 1987). Table S1 in the Supplementary File provides an overview of variables used in the imputation process as well as information on the number and percentage of missing data. Since changes in money management practices were not the focus of our analysis and 50% of the couples in the initial sample were observed only once throughout the survey years, we further restricted our sample to a random draw of one observation per couple and two types of money management. The remaining couple observations were distributed equally across the survey years (Table S2 in the Supplementary File). Our final sample consisted of = 8874 couples.
Our dependent variable was the couple's money management system, which was measured by the survey question: How do you and your partner (or spouse) decide what to do with the income that one or both of you receive? We used three of the five original response categories and recoded them into a dummy variable. Complete pooling (0) was based on the category we pool the money and each take what we need. Independent management of money (1) was measured by the two categories each keeps track of his/her money and we each contribute to a common fund and keep part of our money for ourselves. The latter refers to partial pooling. It requires an independent management to some extent and involves negotiations between partners on spending decisions. We therefore perceived the practice of partial pooling as closer to a completely independent management of money rather than to complete pooling. Even though the stimulus of the question addressed income only, we nevertheless perceived the item to be a valid measure of money management in general due to the broader focus on money in the response categories. Couples with sole management by either the man or the woman were excluded (10% of couples). For most of the couples (87%), both partners' responses were identical. When responses did not match (11%), we randomly selected either the woman's or the man's answer (for a similar approach see Lott, 2017); in 1% of cases, we only had one valid response, which we used as a couple level indicator.
Our independent variables were relative labour income, relative non-housing wealth, and relative housing wealth within the couple. To measure relative income, we used the individual net employment incomes of both partners, which were adjusted for inflation using the Consumer Price Index set to 2015 prices. We focussed on net rather than gross income since it measured disposable income, which can be pooled by the partners or not. After adding up the net income of both partners, we calculated the woman's share of the couple's income and formed four groups: 0-<40% indicated that the woman had a lower income than her partner; 40-<60% indicated income parity within the couple; and 60-100% indicated that the woman had a higher income than her partner; couples without any labour income, i.e., those living on social security entitlements, were coded as a separate category, since some transfers, such as child and housing benefits, were household related and not paid individually.
To calculate relative wealth measures at the couple level, we used the personal gross wealth of both partners, since the gender wealth gap in Germany was mainly driven by gender disparities in gross wealth (Schneebaum et al., 2017) and it better reflected the long-term prospects for asset investments than net wealth. Personal housing wealth referred to the respondent's share of the monetary worth of the owneroccupied property. Personal non-housing wealth was measured by the respondents' shares of other property assets, financial assets, life insurance policies and private pensions, business assets, and tangible assets. These two personal wealth measures were inflationadjusted to 2015 prices and top-coded for the extreme 0.01% of wealth values. After adding up each of the personal wealth measures of both partners, we again calculated the share of couples' housing and non-housing wealth held by women and categorised them into four groups following the operationalisation of labour income. We chose a categorial operationalisation of resource inequalities in order to directly assess possibly gendered associations with money management.
As the overall level of available resources has been shown to affect how couples manage money (Holst & Schupp, 2005;Hu, 2021;Lott, 2009;Ludwig-Mayerhofer, 2006), we added couples' total net household incomes, gross non-housing wealth, and gross housing wealth as controls. Furthermore, we included couples' total debts, comprising mortgages, consumer loans, and other liabilities (again all inflation-adjusted and top-coded). As data on income, wealth, and debts were highly right-skewed, we categorised couples into quartiles for each of these measures. Because there was a high share of couples without either housing wealth or debts, we grouped the first and second quartiles together. To check for multicollinearity, we estimated Cramer's V for the relative and absolute resource measures (Table S3 in the Supplementary File), which indicated low to intermediate associations and thus was not considered to be problematic.
At the couple level, we controlled for further variables that were possible confounders for the association between relative resources within the couple and money management. Compared to married couples, cohabiting couples had higher wealth inequalities (Sierminska et al., 2010) and pooled their money less often (Hiekel et al., 2014;Lott, 2009). The presence of children below age 18 in the household increased the within-couple gender wealth gap  and led to a higher likelihood of pooling money (Hiekel et al., 2014;Lott, 2017). In contrast to the income gap, the withincouple wealth gap remained stable over the course of a partnership (Kapelle & Lersch, 2020), but the likelihood of pooling money increased with the couple duration (Hiekel et al., 2014). The gender wealth gap was significantly higher among West German than among East German couples . Not only did income and wealth gaps vary by education and age (Sierminska et al., 2010), but the likelihood of independent management of money did so as well (Hiekel et al., 2014;Ludwig-Mayerhofer, 2006). Thus, we controlled for both partners' education levels, measured by recoding the CASMIN classification into low (1a, 1b, 2b, 2c_gen), medium (1c, 2a, 2c_voc), and high (3a, 3b) levels of education. Age was measured as age groups for the male partner and the age difference in years within the couple due to the high correlation between both partners' ages. Finally, we added dummies for the survey years to control for time-dependent variations. Summary statistics for all variables can be found in Table S2 in the Supplementary File.
To examine the association between relative resources and the way couples manage money, we first described within-couple resource inequalities in terms of income, non-housing wealth, and housing wealth. Second, we estimated logistic regression models with robust standard errors by applying a pooled cross-sectional design to use the full potential of the available data instead of focussing on one SOEP survey wave as a snapshot. In doing so, we analysed betweencouple differences but not changes within couples. This would have required us to estimate fixed-effects models, which was not possible due to the limited within-couple variation (<15%) in the dependent variable (for a similar approach at the individual level see Lersch, 2017a). Therefore, we were only able to descriptively assess the association between couples' resource constellation and their money management, but not to examine their causal relationship. We presented model results as average marginal effects (AMEs) of the different resource constellations and 95% confidence intervals. We employed a stepwise model setup for each of the relative resources, the corresponding overall resource levels, and the control variables to rule out confounder or suppressor effects (for full regression models see Table S4 in the Supplementary File).
To check that our results were robust, we ran several alternative model specifications and used alternative measurements. First, we applied a more detailed grouping to all relative resource measures (women's share: 0-<20%, 20-<40%, 40-<60%, 60-<80%, 80-100% and couples without income/wealth; see Table S5 in the Supplementary File). Second, we estimated separate models for each survey year with the full sample of couple observations, which, however, meant there were lower case numbers per model (Table S6 in the Supplementary File). Third, we changed the specification of the wealth measures by including overall gross wealth instead of distinguishing between housing and non-housing wealth ( Table S7 in the Supplementary File). Fourth, we used quintiles as an additional specification to measure the overall levels of couples' household income, non-housing wealth, housing wealth, and debt ( Table S8 in the Supplementary File). Finally, we estimated a multinomial logistic model that also included couples with sole money management by one part-ner (Table S9 in the Supplementary File). All alternative model specifications and measurements provided results that were very similar to our main results and indicated no significant deviations.

How Couples' Income and Wealth Constellations Affect Money Management
How do German couples manage money? On average, 68% of German couples in our sample pooled their resources and managed money jointly, while 32% applied independent management (see Table S2 in the Supplementary File). According to our theoretical assumptions, these different money management systems should have been associated with the distribution of financial resources within couples. Figure 1 presents a univariate description of couples' resource constellations regarding labour income, non-housing wealth, and housing wealth. It is apparent that, except for housing wealth, a large proportion of women had lower resources than their male partners. In 60% of couples, the female partner earned less, and, in 45% of couples, the female partner owned less non-housing wealth. In contrast, gender parity was most often reported for housing wealth (43%), but it was evident in only 20% of couples for labour income and 25% of couples for non-housing wealth. Women less frequently earned more and owned more wealth than their male partner, which points towards prevailing gender inequalities in income and wealth within couples and thus supports previous findings (e.g., England et al., 2020;Kapelle & Lersch, 2020;Minkus & Busch-Heizmann, 2020;Nutz & Gritti, 2021). A considerable share of couples did not own residential property (41%), which reflects comparable low  homeownership rates in Germany due to strict mortgage regulations and modest returns to residential investments (Lersch & Dewilde, 2018;Skopek et al., 2012). Couples without any labour income (8%) or non-housing wealth (10%) were not very common.
How were inequalities in these different financial resources related to couples' money management? The results of the pooled logistic regression models for the likelihood of independent money management compared to the pooling system are presented in Figure 2.
Regarding labour income, we expected that an unequal distribution between partners would increase the likelihood of using the pooling system, while withincouple income equality would heighten the probability of employing the independent management system (H1). This assumption was supported by the analyses. The left panel of Figure 2 indicates that income inequalities between partners decreased the probability of using the independent management system. When women earned less than their partners, the couple had a lower probability of managing their money independently by about seven percentage points compared to couples with balanced incomes (model M8). When women had higher earnings than their male partners, the probability was five percentage points lower. This association was strongest when no further controls were applied (model M1) and after considering the couples' overall resource endowment (model M2). Yet it was still significant after controlling for further variables (models M7 and M8).
In line with previous research (Lott, 2009), the likelihood of employing the pooling system thus was higher for couples with unequal labour incomes in both male and female breadwinning couples. It seems that couples with unequal incomes pooled their resources and followed the reciprocal norm of partnership solidarity. The pooling system allows for equal access to money for the partner with the lower income and thus reduces resource inequalities regarding living expenses. In contrast, if income was distributed rather equally between partners, they adhered to the notion of financial autonomy, managed their money independently, and mitigated life risks in the short run rather individually.
Concerning wealth, we anticipated the opposite-we expected that an equal distribution between the partners would make the pooling system more likely, while within-couple wealth inequality would increase the probability of the independent management system (H2). Again, the results supported this assumption. The middle panel of Figure 2 shows that an unequal distribution of non-housing wealth within couples increased the probability of independent money management by about four percentage points, irrespective of whether the women possessed less or more than the male partner (model M8). For housing wealth, this association was even stronger (right panel of Figure 2). If women had a lower share of housing wealth than their partner, this increased the probability of employing an independent management system by 10 percentage points compared to couples with a certain degree of equality in non-housing wealth (model M8). If women owned more housing wealth than their partners, the probability was seven percentage points higher. Again, both associations were strongest in models without any further controls (models M3 and M5) and after considering the couples' respective absolute resource endowment (models M4, M6), but they remained significant after controlling for further variables (models M7 and M8). An unequal wealth distribution within couples, therefore, increased the likelihood of using the independent management system for both housing and non-housing wealth. If both partners held about the same personal wealth, either in terms of non-housing or housing wealth, couples were more likely to pool their money. Couples seemed to follow the norm of partnership solidarity if they owned about an equal share of wealth, typically coming from joint wealth investments. In contrast, persistently unequal wealth distributions within couples increased the likelihood of independent management for both housing and non-housing wealth. This association was evident for both genders. Against the background of within-couple wealth inequalities, couples thus seemed to adhere to the notion of financial autonomy in the long run. This was most obvious for housing wealth, where ownership was gained through buying, donation, or inheritance. But this was also evident for non-housing wealth, which was often owned individually.

Conclusions
In Western industrialised societies, women earn less income and own less wealth than men, particularly within heterosexual couples. While previous research mostly sought to explain why women have lower income and less wealth than men, this article set out to analyse possible consequences of gendered resource constellations within couples. We started with the observation that having access to money in couples was not the same as the ownership of income and wealth but more strongly depended on how couples managed their money. Previous research already indicated that couples' income inequalities mattered in this regard. We expanded this literature by including within-couple wealth inequalities as an additional factor. Accordingly, we asked how couples' income and wealth inequalities were associated with couples' money management. By adding an explicit focus on gendered wealth inequalities, we aimed to contribute to the understanding of wealth as an independent dimension of social stratification (Hällsten & Thaning, 2022;Killewald et al., 2017;Pfeffer & Waitkus, 2021;Skopek et al., 2014).
To understand why couples pool their money or manage it separately, we developed hypotheses based on exchange theory. We perceived these two management systems as representing different norms of reciprocity guiding the exchange process-the norm of partnership solidarity and the norm of individual financial autonomy. Accordingly, we moved beyond the idea of money management as an expression of individual power, as previous studies have done (Lott, 2009;Yodanis & Lauer, 2007). We argued that the money management system adopted has different implications for making up for resource inequalities between partners, and for buffering life risks collectively or individually within the couple.
Our empirical analyses based on data from the SOEP demonstrated that both income and wealth inequalities between partners mattered, yet in opposite directions. Regarding income, a rather equal distribution strengthened the norm of financial autonomy for both partners and increased the likelihood of independent money management. In contrast, if one partner had less income than the other, the pooling of money seemed to strengthen the norm of partnership solidarity, where both gains and losses were shared equally between partners. This was true for both male-and female-breadwinner households, which pointed to historical changes in the way couples manage money: Today, modernized male breadwinning implies a sharing of resources rather than money management by men alone as was the case in the past (Pahl, 1989(Pahl, , 1995Zelizer, 1989). At the same time, resource pooling helps female breadwinning couples to comply with traditional gender roles (Brines, 1994). Overall, our results indicate that money from labour income is easily transferred between partners and can therefore be directly used to buffer financial inequalities within the couple. This is supported by the German modified male breadwinner model, where social policies have been built on the implicit assumptions of separate spheres and of partners sharing their resources.
Regarding wealth, the norms of reciprocity associated with the respective management systems worked in the opposite direction: If both partners held about the same personal wealth, they were more likely to use the pooling system. Accordingly, they followed the norm of partnership solidarity and most likely established this equality through joint wealth investments (Kapelle & Lersch, 2020;Nutz, 2022). In contrast, couples with persisting unequal wealth endowments more often used an independent money management system. Thus, wealth inequalities within couples strengthened the norm of individual financial autonomy. Previous research supports this interpretation, since wealth inequalities within couples only became relevant in certain situations, for instance, when couples separated (Boertien & Lersch, 2021;Kapelle, 2022). Therefore, the wealthier partner had a strong incentive to secure his or her individual long-term financial security, especially if one partner had higher debts (Joseph & Rowlingson, 2012). At the same time, legal wealth regulations often restricted the transfer of assets between partners, which most likely encouraged independent money management. Since men hold more wealth than women, particularly regarding non-housing wealth, this implies that wealth inequalities were not buffered between partners but might have been reinforced by the independent money management system.
Overall, our results demonstrate that money within the household, either in terms of income or wealth, is not a power resource per se as resource theory would suggest. Nor do all couples pool their money to increase the gains of the household as a unit, as the new home economics approach assumes. Rather, couples negotiate how to manage their money based on more complex exchange processes, with particular norms of reciprocity structuring the respective outcomes. About the (re-)production of gender inequalities in general, one major concern was that disparities in economic resources could eventually lead to less pooling and separation of resources between the partners, thereby reproducing inequalities within a couple (Pahl, 1995). However, current research highlights that the couple as a collective unit remains a strong normative ideal. Today, it is a major challenge for couples to balance their commitment to the couple as a collective unit and their pursuit of individual financial autonomy within the partnership (Evertsson & Nyman, 2021;Pepin, 2019). Taken together, our results demonstrate that income and wealth inequalities in couples are relevant for their money management, yet in very different ways. While income inequalities tend to be buffered by a pooling of resources, wealth inequalities are not. Thus, the latter might contribute to a reproduction of gender inequalities also in terms of other outcomes (e.g., Tisch, 2021).
However, the interrelation between these two dimensions of financial resources might matter for the adopted money management strategy too, which may suggest a pathway for further research. Moreover, the adopted money management could also have an impact on couples' future accumulation of wealth, which underlines the importance of a longitudinal analysis. Finally, further research will have to establish whether welfare state policies supporting a dual-earner model compared to male breadwinning result in different associations between partners' wealth constellations and couples' money management. This might also be the case when comparing East and West Germany, where historically varying property and gender regimes resulted in different levels of gender income and wealth inequalities, with possible consequences for couples' money management.
Our analyses faced several limitations. First, the stimulus measuring our dependent variable focussed on income management and not the management of all financial resources or wealth specifically. However, the answers to this stimulus referred to money management in general. It therefore remains an open question whether respondents included wealth management in their answers or not. It is necessary to develop better measures on how couples manage wealth, possibly by differentiating between wealth components. Moreover, the available items in the SOEP survey did not allow us to differentiate between money management, control over money, and spending decisions; nor was it possible to measure the varying norms of reciprocity related to the different money management systems. Finally, the between-couple analysis only allowed us to investigate associations between couples' resource constellations and their money management but not to assess within-couple variations, which would pave the way to a more causal analysis. More fine-grained measurements on how couples manage, control, and spend money would possibly increase the variation of couples' practices and therefore allow more causal analyses of the consequences of gendered income and wealth inequalities.