Abstract
The relationship between an individual’s financial behavior and financial satisfaction is well known. Less evidence is available about how these two constructs interplay within couples. This considered, the current paper aims to (a) examine whether individuals’ financial satisfaction is influenced by their own financial behavior (actor effect) and their partner’s financial behavior (partner effect); (b) examine whether these two effects vary between husbands and wives; and (c) verify how couples’ bank account status (i.e., only joint bank accounts, only separate bank accounts, both joint and separate bank accounts) moderate these effects. The current study draws 1,475 heterosexual early married couples from Couple Relationships and Transition Experiences study and modeled dyadic data through an Actor Partner Interdependence Model. Results indicate that actor’s financial behavior is associated only with one’s own financial satisfaction (actor effect) and not one’s partner’s financial satisfaction (partner effect). This holds for both wives and husbands. Furthermore, individuals who hold only joint bank account(s) are more likely to have financial behaviors similar to their partner than individuals who hold only separate bank accounts or both joint and separate accounts. Couples who hold only separate accounts are more likely to engage in less positive financial behavior than their counterparts. Implications for relationship therapists and financial professionals are discussed.
Keywords
Introduction
One of the first big financial decisions a couple makes when they decide to commit to a long-term romantic relationship, especially marriage, is the type of bank account status they want to have: yours, mine, or ours. In other words, should they keep money in separate accounts? Should they pool their money together in joint accounts? Or, should they have a combination of both separate and joint accounts? Bank account status within marriage has been an area of study for decades. Consider Pahl’s (1980) research on the whole wage system, the allowance system, and the pooling system. Pahl found that couples choose systems that are linked to their family life cycle stage, household income level, and the household’s occupational, regional, and ethnic culture. As couple dynamics have changed throughout the years, with women increasingly participating in the workforce and becoming breadwinners for their households, how couples allocate or manage their money has also changed (Çineli, 2020; Khamis & Ayuso, 2022). Whereas joint or pooling systems were once the primary way in which couples held their money, now couples tend to manage their money more separately or with a combination of joint and separate accounts within their relationships (Lott, 2017).
Although egalitarian type relationships tend to move toward more individual or a combination of individual and joint accounts, previous research indicates that holding joint accounts and experiencing positive relationship outcomes are related (Addo & Sassler, 2010; Olson et al., 2023). While pooling of money (i.e., holding joint accounts) has been a prevalent subject in couples and money research, little attention has been paid to how bank account status (i.e., joint, separate, or combination) impacts dyadic processes around individual and partner financial behavior and financial satisfaction (Garbinsky & Gladstone, 2019; Kruger, 2019). Kruger found that couples who completely combined their finances were significantly more likely to be satisfied financially than those couples who kept their finances separate. Lee and Dustin (2021) found the strongest predictor of financial satisfaction of married individuals is financial behavior (e.g., spending, savings, and planning for retirement). However, they did not look at the dyadic nature of couple relationships. Garbinsky and Gladstone (2019) studied dyadic relationships and discovered that financial behaviors, especially spending, within a couple relationship is influenced by couples’ choice of pooling their money in joint accounts or keeping it separate. If couples pool their money into joint accounts, how do they determine how much one can purchase without consulting the other? If couples choose to keep their money separate, then who pays for which bills, including groceries, mortgage, family car, and other living expenses? For couples who have a combination of joint and separate accounts, which expenses are considered joint expenses, and which are considered individual (i.e., separate) expenses? These are only a few questions that couples must address.
While the literature is sparse in terms of relating couple financial behaviors, bank account status, and financial satisfaction with each other, we adopt a dyadic approach in the current study to examine the interplay of these three concepts (i.e., financial behaviors, bank account status, and financial satisfaction) within marital relationships. The extent to which one’s financial satisfaction is influenced by their own financial behavior or their partner’s financial behavior is not known. Additionally, it is not known the extent that a couple’s bank account status moderates the relationship between financial behavior and financial satisfaction within relationships. Therefore, we use dyadic data to examine variability in financial behavior and financial satisfaction. Our models will examine the relationship of a wife’s financial behavior on their own financial satisfaction and on their husband’s financial satisfaction and vice versa, including models examining the possible moderation of these relationships by bank account status.
More specifically, using a sample of early married (i.e., married less than five years), heterosexual couples, this study has three primary aims: (a) examine whether individuals’ financial satisfaction is influenced by their own financial behavior and their partner’s financial behavior (i.e., actor and partner effects); (b) examine whether these relationships among financial behavior and financial satisfaction vary between husbands and wives (i.e., indistinguishability); and (c) verify how couples’ bank account status (i.e., only joint bank accounts, only separate bank accounts, both joint and separate bank accounts) moderate these relationships between financial behavior and financial satisfaction (i.e., moderation of bank account status).
Relationship between individual financial behavior and financial satisfaction
Financial behavior can be defined as what a person does to manage their personal finances (Xiao, 2008). Positive financial behaviors are exhibited with good planning, management, and financial control activities (Hilgert et al., 2003). Positive financial behaviors (e.g., tracking expenses, budgeting, investing for financial goals) have been hypothesized as constructs that promote financial satisfaction (e.g., Wilmarth et al., 2021).
Financial satisfaction has been defined as being healthy, happy, and worry-free about the individual’s financial condition (Zimmerman, 1995). Similarly, Joo (2008) defined financial satisfaction as indicating if a person’s financial condition is good then they are happy and free from anxiety from their personal financial situation. Given the subjective nature of satisfaction, researchers utilize measures of financial satisfaction based on the respondent’s perception, often using a single item (e.g., Joo & Grable, 2004; Xiao, Chatterjee, & Kim, 2014). There is a body of literature focusing on the promotion of financial satisfaction by emphasizing the improvement of individuals’ financial behaviors (e.g., Hilgert et al., 2003; Hira, 2012; Huston, 2010). Dowling and colleagues (2013) found that lower financial concerns and greater financial satisfaction was a result of having positive financial behaviors. Those who engage in financial management with positive financial behaviors, such as being free of debt, having savings, and paying their bills on time, tend to be more satisfied with their financial condition (Xiao et al., 2006). Financial behaviors that are desirable (positive) increase financial satisfaction, whereas risky (negative) financial behaviors decrease financial satisfaction (Xiao et al., 2009, 2014a).
Gender and finances
When studying the relationship between financial behavior and financial satisfaction at an individual level, inconsistent results regarding gender differences have been found. Because of these inconsistencies, the role gender plays in money and marriage is important to the current investigation. For example, the type of financial behavior being studied is an important factor when looking at possible gender differences. Prior research has identified men as engaging more frequently in credit management behaviors and having more savings than women (Wheeler & Brooks, 2023). Women have been found to engage in more positive financial behaviors (including short-term and long-term behaviors) when compared to men, however they also experience greater financial anxiety and feel less financially secure than their male counterparts (Lind et al., 2020).
In terms of financial satisfaction, previous work has found that there are differences in the way male and female respondents perceive financial issues and that women are typically less financially satisfied then men (Hira & Mugenda, 2000; Xiao, Chen, & Chen, 2014). The main source of financial satisfaction has been identified as financial status for females. For males, financial knowledge was the main source of financial satisfaction (Škreblin Kirbiš et al., 2017). Kulic and colleagues (2020) found that financial satisfaction is associated with money management within a relationship, with gender differences. Women’s financial satisfaction increased when the couple engages in more separate and independent methods of managing their finances. For men, their financial satisfaction increased when they were in charge of financial management for the household, as compared to management by their wife, together, or separate management.
Financial behavior and financial satisfaction in the romantic couple
While many scholars (e.g., Joo, 2008; Joo & Grable, 2004; Lee & Dustin, 2021; Wilmarth et al., 2021) have investigated the relationship between financial behavior and financial satisfaction at the individual level, fewer studies have investigated these constructs within romantic couples. These scientific studies have identified three different kinds of evidence about financial behavior and/or financial satisfaction within couples. First, partners’ financial behaviors can be associated each other. For example, Serido et al. (2015) found that financial behavior of romantic partners is positively associated with one’s own financial behavior among young adults. Second, Bonke and Browning (2009) found that partners’ financial satisfaction are associated, despite husbands and wives often reporting different mean levels of financial satisfaction. Finally, the third interdependent effect corresponds to the association between partner’s financial behavior and one’s financial satisfaction. For example, Kruger (2019) found that respondents who perceived their significant other to be more of a saver than a spender were found to be more likely to be satisfied with their current financial situation compared to respondents who perceived their partner to be more of a spender than a saver.
In sum, we already have evidence that partner’s financial behavior and financial satisfaction are associated each other (e.g., Garbinsky & Gladstone, 2019). The limit of such studies is that they often collect data from just one partner (e.g., Kruger, 2019). To fully study the relationship between financial behavior and financial satisfaction in couples, it is necessary to collect both partners points of view and model their interdependence using an APIM model (Actor-Partner Interdependence Model; Kenny et al., 2006), which considers interdependencies within dyadic data and facilitates statistical modelling of the mutual influence partners may exert on each other. This approach addresses concerns that have been outlined by only using data from one respondent (e.g., Bennett et al., 2010; Vogler et al., 2008). Furthermore, we believe APIM is the right model to test the relationship between financial behavior and financial satisfaction within romantic couples for two reasons.
First, the APIM includes two kinds of effects (Kenny et al., 2006): (a) the actor effect and (b) the partner effect. The actor effect captures how one’s experiences (financial behavior) impacts one’s own outcomes (financial satisfaction). This is what has been usually investigated when the relationship between financial behavior and financial satisfaction has been studied at individual level (e.g., Joo, 2008; Joo & Grable, 2004; Lee & Dustin, 2021; Wilmarth et al., 2021). The partner effect captures how one’s partner’s experiences (partner’s financial behavior) impacts one’s outcomes (ones’ financial satisfaction). This relationship has been investigated in other studies (e.g., Kruger, 2019) but the partner’s experience was reported by the individual who took part in the study and not by the partner itself. APIM has the advantage of testing both actor and partner effects, while also managing the non-independence of data collected from partners of the same couple. Indeed actor and partner effects are estimated while controlling for the other inter-dependent effects within couples: relationship between one’s financial behavior and partner’s financial behavior (e.g., Serido et al., 2015) as well as relationship between one’s financial satisfaction and partner’s financial satisfaction (e.g., Bonke & Browning, 2009). In sum, APIM takes in consideration all the possible effects between financial behavior and satisfaction with couples, when data are reported by both partners.
Second, APIM also allows to test for differences between the two partners (Gistelinck et al., 2018). In particular, it is possible to verify if the actor and the partner effects are invariant across wife and husband (i.e., indistinguishability). This is an important research question for two reasons. First, individual-level research, as reported above, has found some gender differences in the financial domain of life (Hira & Mugenda, 2000). Second, in dyadic-level research as well, some differences have been identified between wives and husbands. For example, Chen et al. (2021), applying an APIM found that when it comes to life satisfaction, wives cared more about their husbands’ absolute income, whereas husbands cared more about their wives’ subjective evaluation of the family’s economic status. Considering this evidence, it becomes important to verify whether and how wives and husbands differently influence each other in the financial domain in order to better understand how financial behavior and financial satisfaction interplay within married couples.
Does the couple bank account(s) status make the difference?
How couples manage their personal finances, including their bank account status, can impact not only financial matters, but also relational constructs within a marriage (LeBaron et al., 2019; Shapiro, 2007). Early work investigated the role that women’s increased labor force participation had on household finances and the role that social and economic factors have within a couple’s financial management (e.g., bank account status; Pahl, 1989). This early work continues to be foundational for the study of couples’ finances (Pahl, 1980, 1989). More recent research continues the work by examining the changes in household structure, women’s increased earning power, and different types of couples (e.g., Addo, 2017; Eroğlu, 2020; Evertsson & Nyman, 2014; Ibragimova & Guseva, 2017; Lott, 2017; Lyssens-Danneboom & Mortelmans, 2014). In a U.S. sample, the majority of the sample thought that couples in romantic relationships (e.g., married or cohabiting) should pool at least some portion of their finances (Pepin, 2019) and in practice, majority of married couples are pooling (completely or partially) their finances (Eickmeyer et al., 2019; Kruger et al., 2023). Even with the majority still pooling their finances, prior research is indicating that having only separate bank accounts and both joint and separate banks accounts are on the rise among couples (Kan & Laurie, 2014; Çineli, 2020).
When finances are managed jointly, there is a greater possibility that each partner in the relationship will have some control over the finances, more likely to have communication around finances, and participate in financial decisions together (Van Raaij et al., 2020). Couples with joint bank accounts tend to experience less financial conflict than those who don’t have joint bank accounts (Duvander & Kridahl, 2021; Lim & Morgan, 2021). Holding a joint bank account was also associated with greater relationship quality for couples, however this association is more consistent for women than it is for men (Addo & Sassler, 2010). The use of joint financial practices provides couples with a greater opportunity to learn more about each other’s finances and financial behaviors (Addo, 2017). Couples who have differing opinions on spending and financial behaviors are more likely to engage in having separate financial accounts (Kruger et al., 2023). While Kruger et al. (2023) did not specifically study financial behavior, their finding on differences in spending patterns provides a foundation for further investigating financial behavior within couples as it relates to bank account status. Additionally, Kruger (2019) found that individuals who reported pooling their income in joint accounts were more likely to be financially satisfied. This study controlled for several financial behaviors, but the behaviors were not specifically investigated in relation to bank account ownership within a couple. While the interplay between financial behavior and financial satisfaction has been well-researched, the role of bank account ownership in the relationship between financial behavior and financial satisfaction is still limited (e.g., Archuleta, 2013; Kruger, 2019). Based on the prior literature, we believe that investigating the possible moderating role of bank account status for the relationship between financial behavior and financial satisfaction within couples, using dyadic data, will further our understanding of the topic. For this reason, in the current study the APIM model described above was also tested separately for three groups of couples based on their bank account status (i.e., only joint bank accounts, only separate bank accounts, both joint and separate bank accounts).
Theoretical framework
We utilized Couples and Finances Theory (CFT) (Archuleta, 2008; Archuleta & Burr, 2015; Archuleta et al., 2013), as a framework of the interdependent relationships among couples and money, also known as relational finances to guide our study. In the most recent version of the framework (Archuleta & Burr, 2015), CFT suggests that relational finances are comprised of the couple relationship system and the financial process system. The couple relationship system is comprised of couple relationship characteristics and relationship or marital quality. The financial process system consists of financial inputs, financial management practices (i.e., financial behavior), and financial well-being (i.e., financial satisfaction). The constructs in both the couple relationship system and the financial process are seen as interdependent rather than linear. According to CFT, the couple system and the financial process system are influenced by individual partner characteristics and larger family ecological contexts (e.g., culture, community, religion, socioeconomic status). CFT is typically used to look at the connections among individual partner characteristics and financial behavior or financial well-being and relationship characteristics or relational well-being (Archuleta, 2013; Baisden et al., 2018; LeBaron-Black et al., 2022; Wheeler & Brooks, 2023). Because both partners as individuals are considered to contribute to the couple relationship and the financial process systems, CFT offers a helpful lens when looking specifically at the interdependent relationships among partners as part of the financial process system (i.e., financial behavior, bank account status, and financial satisfaction; Archuleta & Burr, 2015).
CFT offers propositions that helped motivate the aims and hypotheses of this study, along with previous research. In particular, as stated above, our first aim was to examine whether individuals’ financial satisfaction is influenced by their own financial behavior and their partner’s financial behavior. Regarding this aim, we hypothesized that:
An individual’s financial satisfaction is influenced by ones’ own financial behavior (actor effect).
An individual’s financial satisfaction is influenced by their partner’s financial behavior (partner effect). Our second aim was to examine whether or not the relationships among financial behavior and financial satisfaction within couples (i.e., actor effect and partner effect) vary between husbands and wives. Previous studies have found that the mean level of these variables change across husband and wife, with husbands reporting both higher levels of specific kinds of financial behavior (e.g., saving; Wheeler & Brooks, 2023) and financial satisfaction (Hira & Mugenda, 2000; Xiao, Chen, & Chen, 2014). At the same time, no previous studies have specifically investigated whether the relationships between financial behavior and financial satisfaction (i.e., actor effect and partner effect) vary across husbands and wives, so we have not proposed a specific hypothesis about this. Finally, we aimed to verify how couples’ bank account status (i.e., only joint bank accounts, only separate bank accounts, both joint and separate bank accounts) moderates the relationship between financial behavior and financial satisfaction (i.e., actor effect and partner effect). Previous studies have found that the mean level of these variables change across different couple bank account statuses, with couples holding a joint bank account reporting better financial outcomes (e.g., financial satisfaction; Kruger, 2019) and relational outcomes (e.g., relationship quality; Addo & Sassler, 2010). At the same time, no previous studies have specifically investigated if the relationships between financial behavior and financial satisfaction (i.e., actor effect and partner effect) vary across different couple bank account statuses, so we have not proposed a specific hypothesis about this.
Method
Participants and procedure
Respondents were taken from the Couple Relationships and Transition Experiences (CREATE) study. The CREATE study is a U.S. nationally representative survey of newly married, young couples. The study was approved by all appropriate Institutional Review Boards. Respondents for the study were recruited using a two-stage cluster stratification sample design, with the first stage involving a sample of counties, and the second involving a sample of recent marriages within those selected counties. To be included in the sample, respondents had to (a) be married and selected into the sample frame (since some marriage applicants did not end up marrying), (b) have at least one partner between 18 and 36 years of age at the start of the study, (c) be in a first marriage for at least one of the partners in the dyad, and (d) be living within the U.S.
This study involved 2,181 couples in 2016 (Wave 1) and longitudinally collected data from the same couples in 2017 (Wave 2), in 2018 (Wave 3), in 2019 (Wave 4), in 2020 (COVID Wave), in 2021 (Wave 5), and in 2022 (Wave 6). A seventh wave of data collection is currently ongoing. After Wave 5, four couples requested to be removed from the study, bringing the study sample to 2,177 couples. For the present analysis, we are adopting only Wave 3 data because one of the main variables of interest (i.e. couple bank account(s) status) was only available in Wave 3. See James et al. (2022) for additional information regarding the sampling procedures, recruitment strategies, and response rates.
Data collection for Wave 3 started on May 5th, 2018 and reached 1,722 couples (for 10.4% of couples only one of the two partners completed the survey). For the current study, we included 1,481 of those couples in order to properly investigate role differences (i.e., wife vs. husband); in particular, we removed 57 same-gender couples and 184 couples for which we were not sure the couple was composed by a female and a male (e.g., couples where one of the partner reported “other” as gender or did not fill the gender question). Finally, we removed another six couples where neither of the two partners reported information for any of the variables adopted in the current study (seven items of the financial behavior scale, one item of financial satisfaction). It is worth noting that, having selected only some of the original 2,177 couples of the CREATE study, the sample is no longer fully representative of the U.S. newly married, young couples.
The final sample is composed of 1,475 early married couples where the female partner age range is 21–54 years (M = 31.70; Md = 31.00; SD = 4.73) and the male partner age range is 20–60 years (M = 33.68; Md = 33.00; SD = 5.50). Most of the wives (67.4%) and the husbands (66.9%) were White. Other reported races where Black (6.8% and 9.5%), Asian (5.0% and 3.7%), Native American (.8% and .7%), Latino (12.0% and 12.1%), other (1.5% and 1.3%) or multiracial (6.5% and 5.9%) respectively for wives and husbands. Slightly over 90% (90.7%) of the 1,475 couples were married in 2014, while the remaining couples got married the year before (2013) or after (2015). Most of the husbands (90.6%) and wives (92.7%) were in their first marriage. For 8.5% of husbands and 5.8% of wives this was the second marriage. Finally, for .7% of husbands and 1.5% of wives this was the third marriage. Only three husbands (.2%) had been married more than three times.
Regarding the educational level, more than half of wives had a higher education, in terms of bachelor’s (10.3%), master’s (32.6%), or an advanced degree, such as a PhD (15.1%). The remaining 42.0% of wives had a lower level of education (e.g., high school, some college). Half of husbands in the sample had a higher education, in terms of bachelor’s (10.2%), master’s (27.6%), or an advanced degree, such as a PhD (9.8%). The other half (52.2%) of husbands had a lower level of education.
Regarding the occupational status, most of the wives were working full-time (56.8%) or part-time (12.0%). Few of them were attending school (2.9%) or combining work and school (5.4%). One-fifth of wives (21.5%) were at home full-time. The remaining wives (1.4%) were on official leave. One-third of the wives (28.7%) reported a gross annual household income less than $50,000 (US dollars). Slightly over 40% (41.7%) of the wives reported a gross household income between $50,000 and $100,000, while the remaining 29.6% reported a gross household income of more than $100,000.
Regarding the occupational status, most of the husbands were working full-time (83.6%). Few of them were working part-time (3.8%), attending school (2.0%), or combining work and school (5.5%). The remaining husbands were at home full-time (4.1%) or on official leave (.9%). More than half of the husbands (59.7%) reported a gross annual household income between $50,000 and $100,000 (US dollars). A small percentage of husbands (8.7%) reported less than $50,000 per year gross household income, while a third of them (31.6%) reported more than $100,000 per year.
Measures
Through Qualtrics, respondents filled out an online survey. To study the concepts of interest in the CFT framework, we operationalized those using the following variables from the online survey.
Financial behavior
Each partner’s healthy financial behavior was assessed through seven out of 15 items of Financial Management Behavior Scale (Dew & Xiao, 2011). Two out of the seven items belonged to the “Savings subscale” (e.g., Saved money from every paycheck), three items belonged to the “Cash management subscale” (e.g., Paid all your bills on time), while the remaining two items belonged to the “Credit management subscale” (e.g., Paid off credit card balance in full each month). For each item, the respondent indicated how often s/he engaged in the described behavior in the past six months using the following 5-point Likert scale: 1 = “Never”, 2 = “Seldom”, 3 = “Sometimes”, 4 = “Often”, 5 = “Always”. For each item, respondents also had the option of choosing “Does not apply”. As done in previous studies (e.g., Saxey et al., 2021), we coded responses of “Does not apply” as missing data.
We first verified that this three-factor structure was confirmed both for the wives [χ2 (11) = 32.756, p = .01; RMSEA = .037 (.023, .052); CFI = .989; SRMR = .024] and husbands [χ2 (11) = 35.313, p < .001; RMSEA = .039 (.025, .054); CFI = .989; SRMR = .021]. Despite its multidimensionality, this scale is often used as a unidimensional scale (e.g., Dew & Xiao, 2011; LeBaron et al., 2020; White et al., 2021). We were interested in the partners’ financial behavior, in general, and not in its specific sub-dimensions, the unidimensional model of the scale was tested for both partners using the three subdimensions (i.e., saving, cash management, credit management) as parcels (Little et al., 2002, 2013). Given that this factorial solution was plausible both for wives and husbands, we retained this unidimensional solution and tested the dyadic measurement invariance of the scale (see Table S1 in Online Supplementary Materials) to be sure that the scale conveys the same psychological meaning across informants and that any mean level difference found across partners is a true difference and not a measurement artifact (Tagliabue & Lanz, 2014).
Financial satisfaction
Each partner’s financial satisfaction was assessed through a single item: “Overall, thinking of your assets, debts, and savings, how satisfied are you with your current personal financial condition?” (Sass et al., 2015) rated on a 10-point Likert scale (from 1 = “Not at all satisfied” to 10 = “Extremely satisfied”). Respondents also had the option of choosing “Don’t know” and “Prefer not to say”, which were coded as missing data. The use of a single item to assess the financial satisfaction is common in the literature (for a review see Sorgente & Lanz, 2017). The test of the dyadic measurement invariance is not required for single-item scale.
Couple bank account(s) status
How couples decided to combine or not combine their money in banks was assessed by asking both partners “Which of the following is true of you and your spouse’s bank accounts, including savings, checking, credit cards or line of credit, and investment accounts (except IRAs or Roth IRAs)?” Three response options were available: 1 = “We have only joint bank account(s);” 2 = “We have only separated bank account(s);” and 3 = “We have both joint and separate accounts.” We refer to these three options as: (a) only joint bank account(s), (b) only separate bank account(s), and (c) both joint and separate bank accounts throughout the manuscript.
Data analysis
Descriptive statistics for the main variables of the study were explored using SPSS. Our conceptual model was tested by estimating a series of Structural Equation Models, specifically APIMs, in Mplus, using the Robust Maximum Likelihood (MLR) estimation method.
As suggested by Tagliabue and Donato (2015), missing data were described both at respondent level (i.e., one partner of the couple did not reply to any item adopted in the current study) and at item level (i.e., amount of missing data for a specific item, also including the “Does not apply”, “Don’t know”, and “Prefer not to say” options). Missing data at respondent level were .8% for the female partner and 1.2% for the male partner. Missing at item level, instead, ranged from .9% to 15.9% for the items administered to the female partner, while they ranged from 1.2% to 13.0% for the items administered to the male partner. We addressed missing data using the Full Information Maximum Likelihood (FIML). This method is able to include in the analysis any case that has at least partially complete data. The observed items are used to imply the probable values of missing items, increasing the precision and accuracy of parameter estimates (Enders & Bandalos, 2001).
Aim 1. Actor and partner effects
To test the actor and partner effects between financial behavior and financial satisfaction in married couples, an APIM model was tested using the entire sample. In this model, the financial behavior score was a latent factor for both partners, based on three parcel scores (i.e., savings, cash management, credit management). Furthermore, the financial satisfaction variable for both partners consisted of the standardized scores to have a variable with a variance comparable to that of the financial behavior latent factor (Kline, 2016). Model fit was evaluated using the following indices: RMSEA (Root Mean Square Error of Approximation), SRMR (Standardized Root Mean Square Residual), and CFI (Comparative Fit Index). RMSEA and SRMR lower than .08 are indicative of a sufficient model-data fit, while lower than .05 are indicative of a good model-data fit. CFI higher than .90 is indicative of acceptable fit and CFI higher than .95 is indicative of good fit (Marsh et al., 2004).
Aim 2. Indistinguishability
In order to verify if the described APIM model varied across wife and husband, we performed the indistinguishability test. As suggested by Gistelinck et al. (2018) and previously by Olsen and Kenny (2006), there are six different types of indistinguishability: (a) equal actor effects, (b) equal partner effects, (c) equal residual variances for the outcomes, (d) equal intercepts for the outcomes, (e) equal predictor variances, and (f) equal predictor means. We simultaneously imposed all six parameters to be equivalent for wives and husbands (i.e., “complete indistinguishability (C-IT) test” proposed by Gistelinck et al., 2018). If this constrained model was not different, in terms of model fit, from the freely estimated model tested to address Aim 1, we concluded that the two partners were indistinguishable. As the chi-square test is sensitive to sample size (Kline, 2016), these two models were compared based on their CFI model fit. In particular, the two partners were not considered indistinguishable if the constrained model results in a substantial decrease (higher than .01) of CFI (Chen, 2007). If results indicated that constraints were plausible, these were retained in the next step of analysis.
Aim 3. Moderation of bank account status
Finally, we performed a multi-group analysis and then tested the invariance of model parameters across the three groups to test the moderator effect of bank account status on parameters of the model (Sauer & Dick, 1993; Widaman et al., 2014). In particular, we aimed to verify if the above mentioned APIM model varied across couples who differently deposited their money in bank accounts (i.e., only joint bank account(s) versus only separate bank account(s) versus both joint and separate accounts). As specified in the measures section, both partners reported their couple bank account solution. There were 238 couples where the two partners reported divergent information and 87 couples where only one partner answered the bank account status item. For those 325 couples, we could not calculate the “couple response” on such item. As a consequence, the multi-group analysis was repeated three times: (a) using the wife’s response, (b) using the husband’s response, and (c) using the couple response (by excluding the 325 couples for which it was not possible to aggregate partners’ responses). In the current manuscript we report only results based on the couple perception, while results based on the wife/husband perception are reported in the Online Supplementary Materials (note that results remain the same despite the informant(s) considered). Before running the APIM model separately for the three bank account status groups, we had to verify that the measurement model of the healthy financial behavior latent factor was equivalent across the three groups. In particular, we tested if its factorial structure (i.e., configural measurement invariance), unstandardized factor loadings (i.e., weak measurement invariance), unstandardized intercepts (i.e., strong measurement invariance), and unstandardized residuals (i.e., strict measurement invariance) were equivalent across groups (Dimitrov, 2010). Furthermore, we tested invariance of unstandardized factor variance and covariance as well as invariance of unstandardized factor mean across the three groups (Dimitrov, 2010). If a sufficient level of measurement invariance was met (i.e., at least 80% of items/parcels have invariant factor loadings and intercepts; Dimitrov, 2010), we proceeded testing the invariance of the entire APIM model across the three groups (i.e., only joint bank account(s) versus only separate bank account(s) versus both joint and separate accounts). In accordance with the suggestions of Widaman et al. (2014), we tested the invariance of this structural model following four steps. Specifically, we verified whether model structure (i.e., configural structural invariance), unstandardized beta coefficients (i.e., weak structural invariance), unstandardized intercepts of the dependent variables (i.e., strong structural invariance), and unstandardized residuals of the dependent variables (i.e., strict structural invariance) were equivalent across groups.
Results
Descriptive statistics
Descriptive statistics for financial behavior, financial satisfaction, and bank account status across different informants.
aCouples where the two informants reported divergent information as well as couples where only one partner answered this question were excluded from this count.
Aim 1. Actor and partner effects
We first tested an APIM model, where all the parameters were freely estimated except for the measurement parameters (i.e., parcels’ factor loadings, intercepts, and residuals), which we constrained to be equivalent across wives and husbands when testing the dyadic measurement invariance of the healthy financial behavior scale. This model had very good fit indices [χ2 (20) =103.780, p < .001; RMSEA = .053 (.043, .064); CFI = .972; SRMR = .048] and its results are reported in Figure 1 and Table 2. In particular, we report unstandardized estimates in Figure 1, while standardized estimates are reported in Table 2. APIM on the entire sample (N = 1,475) with parameters free to vary between wives and husbands. Non-standardized values are reported. Standardized parameters of the estimated reported in Figures 1–3.
Findings suggest that, for both partners, one’s financial behavior affects one’s financial satisfaction or, in other words, the actor effects were statistically significant. Regarding the partner effects, we found that, for both partners, partner’s financial behavior did not predict own financial satisfaction. Finally, wives’ financial behavior was significantly associated with husbands’ financial behavior as well as wives’ financial satisfaction was significantly associated with husbands’ financial satisfaction. Taken together, findings suggest that partners’ financial behavior as well as partners’ financial satisfaction are associated with each other, but that actor’s financial behavior is associated only with one’s own financial satisfaction (actor effect) and not one’s partner’s financial satisfaction (partner effect). Hypothesis 1 was supported, while Hypothesis 2 was not supported.
Aim 2. Indistinguishability
We estimated an APIM model where we constrained six parameters (i.e., actor effect, partner effect, outcome residual variances, outcome intercepts, predictor variances, and predictor means) to be equivalent across wives and husbands. This model’s fit indices [χ2 (26) = 121.131, p < .001; RMSEA = .050 (.041, .059); CFI = .968; SRMR = .056] indicated that it was not substantially different from the APIM baseline model (ΔCFI = −.004). We concluded that there are no differences between wives and husbands in any of the six parameters considered and we can treat them as indistinguishable dyads. In other words, regarding our second aim, we found that the way in which one’s financial satisfaction is influenced by ones’ financial behavior and partner’s financial behavior is the same for wife and husband. Findings from this model are represented in Figure 2 (unstandardized estimates) and Table 2 (standardized estimates). We retained these cross-informant constraints in the following step of analysis. APIM on the entire sample (N = 1,475) with parameters equivalent between wives and husbands. Non-standardized values are reported.
Aim 3. Moderation of bank account status
Multigroup measurement invariance for the Healthy Financial Behavior Scale among the three groups.
χ2 = chi-square test; df = degree of freedom; RMSEA = Root Mean Squared Error of Approximation, CI = Confidence Interval; CFI = Comparative Fit Index; SRMR = Standardized Root Mean Square Residual.
We found that the correlation between the financial behavior of the wife with the financial behavior of the husband is higher (r = .895; p < .001) when the two partners only have joint bank accounts(s) than in other conditions (r = .672; p < .001). Successively, when testing equivalence of means, we found that husbands and wives who have only separate bank accounts tend to have less healthy financial behaviors than husbands and wives who have either joint or a combination of joint and separate bank accounts. When constraining these two means (husbands and wives’ levels of financial behavior) to be different from one of the other groups, we retained the constraint that these two means should be equivalent to each other (indistinguishable dyads). In particular, we found that the mean of the “financial behavior” factor can be reduced to −.518 (p < .001) for both partners in the only separate bank account(s) group, while it is equivalent to 0 for partners who have other bank account arrangements.
Multigroup structural invariance of the APIM model across the three groups.
In Figure 3 (unstandardized estimates) and Table 2 (standardized estimates), we reported the parameters’ estimates obtained from the measurement (Table 3) and structural (Table 4) invariance of the APIM model across couples using different bank account arrangements ( i.e., only joint bank account(s) versus only separate bank account(s) versus both joint and separate accounts). APIM on the entire sample (N = 1,156) with parameters equivalent across bank account statuses. Non-standardized values are reported. Dotted lines identified non-invariant parameters, where joint stands for “couples with only joint bank account(s)”, separate stands for “couples with only separate bank account(s)” and both stands for “couples with both joint and separate accounts”.
Discussion
In this study, we set out to understand the interdependency of husband and wives’ financial behavior and financial satisfaction as well as whether this interdependence is moderated by how couples manage their bank accounts (i.e., only joint bank account(s) versus only separate bank account(s) versus both joint and separate accounts). More specifically, we aimed to (a) examine whether individuals’ financial satisfaction is influenced by their own financial behavior (actor effect) and their partner’s financial behavior (partner effect); (b) examine whether these relationships among financial behavior and financial satisfaction vary between husbands and wives (i.e., indistinguishability); and (c) verify how couples’ bank account status (i.e., only joint bank accounts, only separate bank accounts, both joint and separate bank accounts) moderate these relationships between financial behavior and financial satisfaction (i.e., moderation of bank account status). Findings from the current study are discussed separately for each aim.
Aim 1. Actor and partner effects
Our findings suggest that one’s own financial behavior was significantly related to one’s own financial satisfaction (actor effect). However, individuals’ financial behavior was not significantly associated with their partner’s financial satisfaction (partner effect). For example, wives’ reported financial behaviors were significantly associated with wives’ financial satisfaction, but were not significantly associated with their husbands’ financial satisfaction. Likewise, husbands’ financial behaviors were significantly associated with their own financial satisfaction, but not with their wives’ financial satisfaction.
While the partner effect was not significant, our model still confirms that wives’ and husbands’ financial behavior and financial satisfaction are interdependent, as suggested by Couples and Finances Theory. In particular, we found that wives’ financial behavior is significantly associated to husband’s financial behavior as well as wives’ financial satisfaction is significantly associated with husband’s financial satisfaction. This confirms previous studies (Serido et al. (2015) and Bonke & Browning (2009) respectively).
Aim 2. Indistinguishability
Our findings suggest that actor effect and partner effect are invariant across wives and husbands. This means that the way in which one’s financial behavior affects one’s financial satisfaction (actor effect) is the same for wife and husband. Similarly, the way in which partners’ financial behavior (does not) affects one’s financial satisfaction is the same for wife and husband. This result is not in line with previous literature showing that there are differences in the way men and women perceive financial issues (Hira & Mugenda, 2000). We speculate that this inconsistence may depend on the kind of data adopted. Most of the previous literature is based on individual data, where males and females are compared while not being members of the same couples. In the current study, we compare heterosexual partners, that are expected to be more similar to each other than two randomly selected different-gendered individuals. Furthermore, previous studies have focused on differences in the mean level of financial behavior (e.g., Wheeler & Brooks, 2023) and financial satisfaction (e.g., Hira & Mugenda, 2000; Xiao, Chen, & Chen, 2014), while in the current study we tested whether gender made the difference in the way these two constructs are related each other.
Aim 3. Moderation of bank account status
Our findings suggest that the couple bank account status (i.e., only joint bank accounts, only separate bank accounts, both joint and separate bank accounts) does not moderate the way in which financial behavior and financial satisfaction are associated within early married couples. In other words, the way in which ones’ financial behavior affects ones’ financial satisfaction (i.e., actor effect) is the same for couples who adopt only joint versus only separate versus both joint and separate bank accounts. Similarly, the way in which a partner’s financial behavior (does not) affects one’s financial satisfaction (i.e., partner effect) is the same regardless the bank account status of the couple.
At the same time, we found the couple bank account status made the difference in terms of healthy financial behavior for two reasons. First, we found that partners who had joint bank account(s) had more similar financial behaviors with each other. In other words, wives who reported high levels of healthy financial behavior are more likely married with husbands reporting high levels of healthy financial behavior and vice versa. These findings contradict previous research that suggests that opposites attract when it comes to financial behavior—spenders marry savers and vice versa (Britt et al., 2017). However, differences in financial behavior between partners tends to create conflict within the couple relationship. Again, the present study is methodologically different from Britt et al. (2017). We looked only at financial variables (i.e., financial behavior and financial satisfaction) rather than including relational variables (e.g., conflict). We also did not examine how partners perceived the other partner’s financial behaviors. As a result, couples in this study tended to align in terms of how they behaved with their money.
Second, we found that partners who held only separate bank accounts were less likely to engage in positive financial behaviors than their counterparts who held only joint bank account(s) or a combination of joint and separate accounts. This result aligns with Kruger’s (2019) findings that reported couples who agree on spending issues are more likely to hold joint bank accounts. Taken together, these findings infer that couples who hold joint bank account(s), whether it be only joint account(s) or a combo of joint and separate accounts, may be more accountable to someone else (i.e., their partner) for how they individually manage their money. A sense of accountability may lead to an increased effort to pursue positive financial behaviors. While not a variable in the current study, couples who have joint accounts may also have an increased sense of trust in their partner and a belief that working together will lead to increased satisfaction about their financial situation. Other research has found that couples who hold joint accounts are less likely to engage in financial conflict (Duvander & Kridahl, 2021). With this in mind and taking into account the results of the current study, a couple who has only separate accounts may be trying to avert conflict that arises from a difference in how each partner manages their money or their perception of their partner’s financial behavior.
Couples and Finances Theory (Archuleta & Burr, 2015) provides a lens to interpret the dyadic findings, as discussed already. Relational finances help us to understand that the couples’ financial behaviors and bank account status cannot be seen as separate from the couple. Rather couples do things (i.e., behavior) based upon their individual and couple financial and relational beliefs and values (e.g., Should we as a couple manage our money together or separately?). Previous research and CFT suggests that when couples are more in alignment in what they do with their money, they may have increased trust in their partner because they are more likely to see eye-to-eye in how to spend, save, and use credit. Qualitative research has suggested that trust is a factor in jointly managing couples’ finances (Burgoyne et al., 2007; Skogrand et al., 2011). Future research should quantitatively examine CFT concepts dyadically, including couple relationship system variables, like level of trust within the relationship, and the financial process, such as partner’s ability to manage money in relation to bank account status and financial behavior.
Limitations
As with any study, the current research is not without limitations although they were kept at a minimum. First, we used cross-sectional data rather than longitudinal data to examine the interdependency of early married couples (less than five years of marriage). Longitudinal data could have helped to better understand directionality between variables under investigation. Second, we only used data from couples who reported similar bank account ownership status and excluded the data from couples where only one partner reported bank account ownership status or the reported status was not in agreement. However, we did test two other models. The first model tested data using all of the wives and second model in which all husbands were included regardless of if their response was in agreement with their partner. These two models produced similar results as the model reported in this study. Third, a more precise measure of bank account management status could have been used. For example, differences in partners’ responses could have been because partners could have a joint account together but only one partner has a separate account. Fourth, while the purpose of this research was to examine couples who were in the early stage of marriage, only couples in Wave 3 of the data or that had been married around four years were included in the study. As a result, the findings are not generalizable to all couples, especially longer married couples. Fifth, available data were not sufficient to describe the sample about all the possible personal characteristics (e.g., disability) that could make the difference in terms of marriage and finances. Finally, we only examined the financial process system of the larger CFT framework. However, CFT suggests that interdependency exists between couple relationship system concepts and financial processes. Including couple relationship variables may have produced different results. CFT does demonstrate the dependency with individual partners as well, which helps to support the use of dyadic analyses in the present study.
Implications
We believe that many of the findings of the current study has implications for practitioners. First, we confirmed the relationship between one’s financial behavior and one’s financial satisfaction. Financial education programs can work on clients’ individual financial behavior which may have a direct effect on their financial satisfaction.
Second, we found that a partner’s financial behavior does not affect one’s own financial satisfaction. Professionals (e.g., financial therapists, mental health and relationship professionals, and financial professionals) working with couples can encourage partners to be responsible and accountable for their own financial behavior as one’s financial satisfaction depends more on one’s own financial behavior. These professionals can help to restructure how couples talk to each other to reduce blaming their partner for their financial behavior if they are experiencing decreased financial satisfaction.
Third, the results reported above (i.e., one’s financial satisfaction depending only from one’s financial behavior and not from partner’s financial behavior) work similarly for both wife and husband, so practitioners should recognize that previously identified (or perceived) strong differences between males and females in the financial domain are nowadays reduced. Couple dynamics have changed throughout the years with women increasingly participating in the workforce and even becoming breadwinners for their households (Çineli, 2020; Khamis & Ayuso, 2022); this may have reduced the difference across gender when it comes to money and finances.
Finally, findings from this study support the notion that joint bank account(s) status reaps positive outcomes, and that holding separate account(s) can be problematic in terms of engaging in less positive financial behaviors. However, a relationship therapist or financial professional should be cognizant that couples who choose to have joint accounts are in more alignment with what to do with their money. If a couple has differing ways in which they behave with their money and causes distress in the marriage, they should not necessarily hold joint accounts. Rather, the findings suggest that if couples are in alignment with how to manage money they may want to consider holding joint accounts.
Conclusion
In summary, we set out to further understand the interdependent nature of couple relationships and finances. We found that partners’ financial behavior as well as partners’ financial satisfaction are associated with each other, despite partners’ financial behavior not affecting one’s own financial satisfaction. Furthermore, individuals who hold only joint bank account(s) are more likely to have financial behavior similar to the one of the ones’ partner than individuals who hold only separate bank accounts or both joint and separate accounts. Couples who hold only separate account(s) are more likely to engage in less positive financial behavior than their counterparts. These findings begin to shed light on previous research that has indicated that one’s bank account status (i.e., joint or separate) is related to couple conflict. Future studies should replicate our study taking also into consideration relational outcomes, in order to further understand the interplay between financial process and couple relationship systems. For now, when considering “yours, mine, and ours” in regards to finances, “our” money is best when “yours” and “mine” financial behavior are in alignment.
Supplemental Material
Supplemental Material - Yours, mine, or ours: Does bank account status in early marriage affect financial behavior and financial satisfaction?
Supplemental Material for Yours, mine, or ours: Does bank account status in early marriage affect financial behavior and financial satisfaction? by Angela Sorgente, Margherita Lanz, Semira Tagliabue, Melissa J Wilmarth, Kristy L Archuleta, Jeremy Yorgason and Spencer James in Journal of Social and Personal Relationships
Footnotes
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by funding from the School of Family Life at Brigham Young University.
Open research statement
As part of IARR's encouragement of open research practices, the authors have provided the following information: This research was not pre-registered. The data used in the research are not available. The materials used in the research are available. The materials can be obtained by emailing:
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
