Debt Concordance and Relationship Quality: A Couple-Level Analysis

Abstract

Despite a large literature on household finances and relationship quality, little is known about the degree of couple-level agreement on finances and its association with relationship outcomes. This study examines the relationship between financial concordance on household-level consumer debt and relationship quality, and the strength of the association after accounting for couple-level financial management practices. We apply hierarchical linear modeling (HLM) to couple-level data from a sample of married and cohabiting couples with children (435 couples, 870 respondents) from the Marital and Relationship Survey (MARS) and find that concordance on total household credit card debt is common but not the norm, with 55% of couples agreeing on their outstanding debt amount within our sample. Debt concordant couples have greater relationship satisfaction even after accounting for the outstanding debt amount, financial management practices such as income pooling, and joint purchase decisions, as well as relationship characteristics like their marital status, relationship duration, and the number of children in the household. We also find that disagreements related to financial issues attenuate the debt concordance and relationship quality association. Our results highlight the importance of including objective measures of household finance when assessing relationship quality.

Introduction

Research on financial organization and relationship outcomes has primarily focused on the control and management of income. Income pooling and joint bank account practices have been associated with better relationship quality and relationship cohesion (Addo and Sassler 2010). Self-reported indicators of couple-level concordance on financial issues, such as how money is being spent, saved, or managed have been shown to reflect increased communication and reduced conflict (Britt, Huston, and Durband 2010; Papp, Cummings, and Goeke-Morey 2009; Skogrand et al. 2011). They can also serve as indicators of better relationship quality and satisfaction, as well as signal lower likelihoods of marital disruption or divorce (Halpern-Meekin and Tach 2013; Totenhagen et al. 2019). Subjective measures of financial arrangements and behaviors, however, can bias estimations if individuals associate financial behaviors with relationship outcomes (Addo and Sassler 2010; Woolley 2003).

Less attention has been given to objective indicators of agreement on financial issues between couples. This is surprising given the plethora of research on joint decision-making on household purchases, asymmetry in who handles household finances, and evidence regarding the lack of intra-couple communication when it comes to financial issues (Breunig et al. 2005). While these may reflect to an extent the degree of understanding of the household’s financial status between partners, studies find problems may arise if partners are not be completely truthful with one another on the actual state of their financial affairs or honest about their satisfaction with their financial situation (Archuleta et al. 2011). Conflict related to household finances can prove to be particularly challenging for couples attempting to maintain stable and supportive unions. Not to mention, structural issues related to financial management, such as the ability to purchase or charge items quickly and independent of consensus, may also contribute to couples disagreeing on their finances and associated with intrarelationship discontent. These issues may become particularly salient when examining negative household assets, like consumer debt. Carrying lots of credit card debt can be stigmatizing and may unfairly connote poor financial management skills (Durkin 2000; Mann 2008). As access to consumer credit and the accumulation of consumer debt have become increasingly common across the income distribution, there is growing theoretical and empirical interest on the role of debt and debt management for families and households.

This study delves into the black box of intrahousehold finance to shed light on the role of financial concordance on relationship quality. Concordance in relationship practices is indicative of agreed upon steps, communication, and joint decision-making of the relationship. Concordance can signify commitment to the couple, which is important for securing long-term marital and relationship stability (Stanley, Rhoades, and Whitton 2010a). Using couple-level data from the Marital and Relationship Study (MARS), an online relationship survey of cohabiting and married couples with small children, we analyzed the degree to which couples agreed on their outstanding consumer debt by exploiting survey data provided by both members of a couple collected separately. We then applied hierarchical linear modeling (HLM) to our dyadic data to analyze the association between consumer debt concordance and relationship quality. This study differs from prior work on consumer debt and relationship outcomes (Dew 2007; Schaninger and Buss 1986) in that we focus on whether couples agreed on the amount of household debt, after accounting for the amount of consumer debt, and investigate whether agreement was correlated with the quality of their relationship. Our study draws from and builds upon family demographic research that finds that concordance in relationship attributes such as relationship tempo and intent matter for relationship commitment, cohesion, and well-being (Brown and Booth 1996; Halpern-Meekin and Tach 2013; Sassler, Addo, and Lichter 2013) and the relevance of household debt for couple dynamics and outcomes.

Background

Household Finances and Relationship Quality

Most individuals within families and households often engage in common practices related to the allocation and distribution of shared resources. Couples who tend to practice joint management skills and share control over financial decisions are also more likely to share other financial practices (Woolley 2003) . These tasks are accomplished often by engaging in practices that improve individual and couple-level well-being through reduced transaction costs and increased efficiency (Halpern-Meekin and Tach 2013; Pollak 1985; Sassler, Addo, and Lichter 2013). Specific to the household finances, studies find that couples’ decisions regarding how to administer their income – whether one partner manages the financial resources (and which partner it is), if they pool their income or maintain separate pots, or hold joint or separate bank accounts – indicates the level of investment and integration in a relationship (Addo 2017; Heimdal and Houseknecht 2003; Oropesa, Landale, and Kenkre 2003; Pahl 2008; Treas 1993).

Jointly managing household finances and making mutual financial decisions becomes a means of signifying and solidifying commitment to the couple, which is important for securing long-term marital and relationship stability (Stanley, Rhoades, and Whitton 2010b). Financial organization and management can serve as proxies for long-term investments in the relationship or indicate a degree of individuality or distrust that could reduce relationship quality. Stanley et al. (2010) assert that relationship practices that increase the costs of relationship dissolution can positively influence the overall stability of the relationship independent of romantic attachment. Pooling systems, such as combining incomes in households or sharing management over financial decisions, has been shown to reflect better relationship quality. The collective orientation demonstrated by those who share fiscal management and organization may also reduce conflict. Perceived shared values on finances among couples, such as having shared financial values and making financial decisions together, are linked to improved marital relationship satisfaction (Totenhagen et al. 2019).

Studies on household finances and relationship outcomes primarily focus on the organization and management of finances. It is not empirically clear if practices that are related to household financial management reflect agreement about household finances, and less studied and understood is whether couples are in agreement about the current state of their finances. Nor has it been shown, to our knowledge, that income pooling, joint decision-making, compromising, and lower conflict about finances serve as potential proxies for financial concordance between partners. There is reason to believe that financial concordance may be an important independent indicator of relationship dynamics within the household. We believe this may be the case for at least two reasons. First, there are opportunities for partners to withhold information from each other or circumvent the management system. Second, the potential for discordance is related to the ambiguities with the financial organization process, and not necessarily reflective of the couple’s attempt at financial transparency. We describe in the next section how discordance in finances can arise in the particular case of consumer debt related to these two scenarios.

Consumer Debt, Household Finance, and Relationship Quality

Household financial practices related to debt, and credit management have increasingly received attention by family scholars. Research on household behavior and management of consumer credit had historically focused on married households and treated debt as a household-level asset. Prior to 1974 and the passage of the Equal Credit Opportunity Act (ECOA), all charges, i.e. debt, went into a joint household account under the husband’s name (Blakely 1981). After the ECOA, not only were creditors no longer able to discriminate against female applicants, wives were also able to obtain credit accounts separate from their husbands. Credit cards act as a “democratizing” agent allowing individuals with access across the economic distribution to purchase and consume goods in the marketplace and delay repayment to a later period at almost always a higher cost (Mann 2008). They increasingly became a mechanism for making small (e.g. groceries, paying bills) to large purchases (e.g. car payments, rent), with households charging everything from medical bills, cars, to education loans (Draut and Silva 2003). Despite the short-run convenience of instantaneous credit access, market incentives to roll over debt from one period to the next rather than pay it off completely are quite high and is associated with debt accrual and ballooning debt burdens. When combined with variable interest rates, outstanding debt becomes increasingly expensive and hard to get rid of due to compound interest (Durkin 2000).

Low-income populations, in particular, have experienced significant growth in credit card utilization and credit card-based financial debt burdens (Bird, Hagstrom, and Wild 1999; Lyons 2003; Mann 2009). Scholars also point to the relative rise of credit card usage during a period of decreasing social safety nets (Mann 2009). Based on data from the Survey of Consumer Finance, Scholz and Seshadri (2009) found that from 1989 to 2004, positive credit card debt holdings increased the most for those in the bottom two income quintiles, by about 12–13%. By 2006, 30% of low-income households and 43% of moderate-income households held positive credit card debt, both very close to the overall U.S. average of 46% (Mann 2009).

Understanding credit card and debt behavior can be challenging, as credit spending behaviors may signify different or conflicting financial patterns. For example, having no credit cards can indicate households that have an aversion to holding credit or being in debt. Alternatively, it can be a signal of prior credit problems or the inability to access credit card markets, which is more likely among low-income populations and those who are financially vulnerable. People can also actively chose not to use credit cards to make purchases, relying instead on cash, checks, or alternative inferior financial systems such as payday loans, which are used more by lower income populations (Barr 2004). At the household-level, this can be confounded if spending habits and money management styles differ between partners (Kirchler, Hoelzl, and Kamleitner 2008; Rick, Small, and Finkel 2011). If one partner is debt averse, but the other is not, they may be less likely to agree on financing purchases with credit cards or how to pay off a particular debt balance.

Similar to bank accounts, potential borrowers can apply for and hold a joint credit accounts with a partner in addition to their own separate account. This affords partners the ability to fulfill a collectivist and independent financial agent roles within relationships (Pahl 2008). In contrast to bank accounts that predominantly provide access to savings, such as positive income and savings, combined credit card accounts link account holders through the accumulation of debt, a negative household asset and a shared credit history. Studies that have examined the role of outstanding debt amounts find that it is positively associated with relationship conflict (Dew 2007) and increases the probability of divorce (Dew 2011). With separate accounts, unsecured consumer debt can be accumulated and managed without the knowledge of one’s partner (Addo 2017; Pahl 2008). And with combined accounts, separate cards can indirectly contribute to couple-level discordance related to outstanding debt balances and debt behaviors. This structural feature of credit accounts and card accounting may introduce ambiguity into the household debt accumulation process.

Finally, debt is a continuous measure. Oftentimes when asked in surveys about monetary figures, respondents may not be able to recall exact dollar amounts or misremember. They are even more inclined to be myopic about their purchasing habits, are more likely to forget the costs of large items that are purchased infrequently, and more inclined to remember smaller purchases that are consumed more regularly (Hurd and Rohwedder 2009). In addition, the convenience of the credit transaction process increases the likelihood of use. However, as technological advances related to card payment and credit account access have made it more convenient to charge purchases, it has also facilitated the capability to check balances and stay informed of one’s outstanding debts on a regular basis, possibly making it easier to keep track of household finances (Kolodinsky, Hogarth, and Hilgert 2004).

The Current Study

In the current study, we are interested in whether an objective indicator of financial cohesion, agreement in household credit card debt between partners, is correlated with relationship quality. Concordance in relationship practices is indicative of agreed upon steps, communication, and joint decision-making of the relationship. Does the transactional translate into the relational? We argue that concordance is reflective of relationship transparency and an investment in efficient and transaction cost-reducing practices that may confer better relationship satisfaction. We begin with an exploration of the prevalence of credit card debt, debt behaviors, and debt concordance within our sample, and in addition investigate how several joint financial relationship practices vary by debt concordance status. We hypothesize that collectivist financial practices will be more prevalent among debt concordant couples (Hypothesis 1). We then turn to the concordance and quality relationship and posit that debt concordance will be positively associated with relationship quality with concordant partners reporting higher overall relationship satisfaction (Hypothesis 2).

We also anticipate that the strength of the relationship between debt concordance and relationship quality will be attenuated by several relationship and financial management characteristics (Hypothesis 3). Debt behaviors and financial management can vary by relationship type and the rituals surrounding relationship formation. Relative to married couples, cohabiting couples are less likely to combine finances, such as having joint bank accounts (Addo and Sassler 2010; Bellah et al. 1985; Heimdal and Houseknecht 2003). Cohabiting and married couples with children, however, practice similar money management styles compared to couples without children (Vogler, Brockmann, and Wiggins 2008). Studies of married and cohabiting couples find that relationship quality declines over time (Brown 2003; Musick and Bumpass 2012; Umberson et al. 2005). In addition, prior marriage experience and having children have been associated with decreased relationship quality (Lerman 2002; Treas 1993; Waite and Lillard 1991; Woolley 2003). Cohabiting partners are also much less likely to recall the progression of their relationship similarly, which is associated with decreased relationship quality (Halpern-Meekin and Tach 2013).

The relationship quality of couples who decide to formally enter in a relationship, such as marriage, differ from cohabiting couples, due to the degree of commitment that manifests themselves differently across these relationship types (Stanley, Rhoades, and Whitton 2010b). Commitment consists of two parts: personal dedication, which secures romantic attachments, and constraints, which is what leads to action or inaction from one relationship phase to the next. Dedication increases the likelihood that couples have better relationship quality, whereas constraints influence the overall stability of the relationship, independent of romantic attachment. Stanley et al. (2010) argue that for some married couples who cohabited prior to marriage, relationship inertia kept them together. As the number of constraints such as integrated money management systems increases, it may be difficult to separate as a couple (Addo 2017). We therefore also expect that couples who pool their income (Pahl 1995), compared to having money completely separate or having one partner manage all the money, are more likely to be debt concordant (Hypothesis 4).

Information about financial organization and indicators of financial cohesion may also influence differences in relationship quality related to debt concordance. They can serve as proxies for long-term investment in the relationship or indicate a degree of individuality that could reduce relationship quality. Research on who pays the bills and makes household purchase decisions has found asymmetrical gender differences, with women in charge of bills not necessarily because they were decision-makers but due to being viewed as household homemaker duties. When it came to decision-making on large-scale purchases, however, these were often made by the husband or through consensus (Bennett 2013; Pahl 2008). Pooling systems in which the household combined income and shared management over financial decisions has been shown to reflect better relationship quality, for example (Addo and Sassler 2010; Kenney 2006). Control over how finances are spent should also be positively correlated with debt concordance if the couple practices joint decision-making. Couples who tend to practice joint management skills and share control over financial decisions are also more likely to share other financial practices (Woolley 2003). The collective orientation demonstrated by those who share fiscal management decisions may have lower conflict and report better conflict resolution all manifesting in better quality relationships.

Additional factors that may influence relationship quality include conflict over management of money and power imbalances within the relationship. Couples may argue over how they spend their money, which may lead to additional strain within the relationship if both partners disagree with spending habits, thus decreasing relationship satisfaction (Archuleta et al. 2011; Papp, Cummings, and Goeke-Morey 2009). And while not necessarily the source of relationship dissolution, disagreement on financial issues is commonly cited as a reason for relationship discord (Papp, Cummings, and Goeke-Morey 2009). This can manifest in negative perceptions of one’s partner and the relationship (Conger, Rueter, and Elder 1999). Partners who perceive power imbalance within the relationship are also more likely to report lower relationship satisfaction, and couples who compromise are more likely to report higher relationship satisfaction (Pistole 1989; Sprecher and Felmlee 1997).

Several additional sociodemographic and relationship characteristics may also influence the relationship between debt concordance and relationship quality, such as the respondent’s race or ethnicity. Black couples have been more likely to report lower relationship quality relative to white and Latinx couples (Bulanda and Brown 2007), but the explanation of differences by race is unknown. Educational attainment has also been linked to relationship quality, with those who are more highly educated being less likely to divorce or report poorer relationship quality (Brown 2003; Brown and Booth 1996; Glenn 1990). Economic stability, such as being employed and having greater household income, has been associated with positive relationship quality, while experiencing material or economic hardships has been associated with lower relationship quality (Conger et al. 1990; Lichter and Carmalt 2009). Couples who face financial constraints may respond in two different directions: they may be well-suited to have better relationship commitment due to adapting to stress positively (LeBaron et al. 2020), or have increased relational aggression, or reduced support towards the partner (Wheeler, Kerpelman, and Yorgason 2019). These individual and relationship attributes are also included as additional covariates, in order to isolate the main relationship of interest, the association between debt concordance and relationship quality.

Data

We used data from the Marital and Relationship Survey (MARS), an internet-based survey administered once in 2006 by Knowledge Networks (KN)¹. KN maintains a nationally representative web-enabled panel of respondents of the online and offline population in the United States. Panelists receive a unique login and password to ensure confidentiality of responses. The MARS sample was selected from the KN panel of respondents and consisted of married and cohabiting couples with coresidential children. Designed to be representative of the targeted population for marriage promotion policies after the 1996 welfare reform was renewed in 2006, the household incomes of MARS respondents had to be less than $50,000 and the female partner had to be between the ages of 15–44. Surveys were administered to each partnerin separate interview sessions and took approximately 35–40 minutes to complete. Given that KN panel members were selected at random and contractually agreed to complete the survey, self-selection bias and non-response error were minimized, and the MARS response rate was 80.3%. At the time of interview, several households did not have both partners available to participate and only one person was interviewed. As a result, we eliminated 122 single-person households from the original 1,095 interviewed. An additional 100 respondents (50 couples) were not included in our sample because either one or both of them did not answer questions on household debt, and one couple because a partner did not answer the relationship satisfaction question. Our final analytic sample consisted of 870 respondents from 435 couples. Though not a criterion for inclusion, all couples were heterosexual.

Measures

Relationship satisfaction.

Respondents were asked to assess their satisfaction with their romantic relationships ranging from 0 (not at all satisfied) to 10 (completely satisfied).

Debt concordance.

The debt concordance measure was coded based on the intra-couple level responses to the following question: “What is your combined credit card balance?” Respondents could either provide specific dollar amounts including zero or specify that they had no debt either because they paid off the balance on their credit card every month or because they did not own a credit card. Couples are defined as concordant if they both reported the exact same amount of credit card debt, or if they both reported not having debt for the same reason (e.g. both partners reported zero debt because they had no credit cards), and discordant otherwise.

Relationship characteristics.

We included several characteristics of the current relationship, such as whether the couple was currently cohabiting or married. Married respondents were disaggregated into those who cohabited prior to marriage and those who directly married their current spouse without cohabiting beforehand. We included an indicator variable for having been previously married, a continuous measure of relationship duration in months, and the total number of children ages 18 or younger present in the household. We also included a measure related to conflict resolution. Respondents were asked to specify who usually gets their way when disagreements arise between them and their partner. An indicator variable was created for partners that specified that they generally compromise equal to one, with responses that either the spouse or the respondent gets their way equal to zero.

Couple-level financial management practices.

We included four measures to assess varying aspects of couple-level financial management practices. The first measure was income management, equal to one if the respondent selected that they either “pool all the money and each take out what we need” or “we pool some of the money and keep the rest separate” in response to which statement best describes how you and your [spouse/partner] organize the income that one or both of you receive, and zero otherwise. The second measure was on couple-level decision making, which was equal to one if the respondent selected that “we decide together” for who usually makes the decision on big purchases (e.g. cars, expensive appliances) and zero otherwise. The third measure was whether the respondent held a joint bank account with their partner, equal to one if they did and zero otherwise. Our final measure captured the frequency of arguments related to managing money, bills and debt. Respondents could select from never, a few times a year, or few times a month, or a few times a week or more. Given the uniformity in the distribution of responses we grouped the responses for a few times a month and few times a week or more together creating three distinct categories.

Socioeconomic measures.

We also included a host of socioeconomic characteristics that have been shown to be associated with relationship quality. These consisted of the respondent’s educational attainment, with the categories of less than high school, high school (reference category), some college, and college graduate or more, their current employment status (employed full-time, coded as 1 or not, coded as 0), and a logged measure of total household income from all wages from salary or self-employment. The financial status of the household was from a continuous count of the number of economic or housing quality-related material hardships experienced at any point during the previous calendar year and a measure of the outstanding consumer debt amount. In addition, we included the amount of credit card debt and logged the non-zero values in the model.

Demographic measures.

In addition to the respondent’s gender (male or female), our main race and ethnicity categories are non-Latinx White (the reference category), non-Latinx Black, and Latinx.

Methods

We begin by presenting basic descriptive statistics of our sample. We then explore the financial organization and management measures with a focus on differences by concordance status. Our final stage of analysis presents empirical estimates for our relationship satisfaction models. Hierarchical Linear Models (HLM) using maximum likelihood estimation was used (Raudenbush and Bryk 1992) and appropriate given the dyadic structure of the dataset, which included individuals (Level 1) nested within couples (Level 2). This nested design means that observations related to individual partners within the sample couple were interdependent, violating OLS assumptions of non-independence of observations. HLM relaxes this assumption allowing for correlated error structures (Luke 2004). For each model, the intercept was allowed to vary randomly across couples and all continuous variables were centered at the grand mean. We used Stata 16.0 to estimate our models using the mixed command. The analyses proceed in four steps. First, we examined the unconditional HLM, which computed the average relationship satisfaction with the sample. We then have the conditional models that include dyadic predictors and individual-level control variables. Model 2 includes relationship characteristics, Model 3 adds in financial management practices, and Model 4 adjust for the frequency of arguments related to finances. With the exceptions of the debt measure (5.5% missing) and household income (8.8%), missing data was less than 2% for all model covariates. Missing data was multiply imputed by chained equations using the mi chained command in Stata (Royston and White 2011). We used the “multiple imputation, then deletion” (MID) method in which respondents with missing information on the dependent variable were included in the imputation, but not in the final analytic sample (Von Hippel 2007). The final analysis is based on five implicates of imputed data.

Results

Sample Descriptives

Table 1 shows descriptive statistics for the full sample. The majority of the sample (90%) was married, with the remaining 10% in cohabiting unions. Fifty-four percent of the married respondents reported living with their partner prior to marrying. Couples had about two children on average and 22% had been previously married. The average length of relationships was approximately 122 months (10 years), an indication that the respondent pool was skewed towards couples in long-term relationships. With regards to relationship characteristics, almost two-thirds (63%) of the sample reported that they compromised during disagreements. Most couples reported that they argued at least a few times a year over managing money, bills, and debt (74%).

Table 1.

Analytic Sample Summary Statistics (N=870)

	Mean+	Std. Dev.
A. Relationship Characteristics
Current Relationship Status
Direct Marriage	0.41
Cohabited, then Married	0.49
Currently Cohabiting	0.10
Total Number of Children in Household (1–7)	2.08	1.11
Previously Married	0.22
Relationship Length in months	122.07	2.42
Compromise during arguments	0.62	0.48
B. Demographic and Socioeconomic Characteristics
Female	0.50
Race/Ethnicity
Non-Latino White	0.88
Non-Latino Black	0.05
Latino	0.07
Educational Attainment
Less than High School	0.10
High School	0.33
Some College	0.38
Bachelors or more	0.19
Currently employed full-time	0.52
Household Income (logged)	8.15	3.82
Number of economic/material hardships (0–10)	2.15	2.16
C. Outstanding Credit Card Debt/Debt Behavior
Non zero outstanding debt	0.58
Credit card balance (logged)	3.86	0.14
No balance carried	0.17
No credit card	0.25
Consumer debt concordance	0.55
Concordance on:*
Outstanding debt amount	0.41
Zero balance, paid off monthly	0.22
No credit cards	0.37
D. Relationship Satisfaction
Relationship Satisfaction (0–10)	8.32	1.77
Women	8.19	0.09
Men	8.46	0.08
Within couple difference	1.10	0.05

Note:

⁺Mean and standard deviation are reported for continuous measures and proportions for binary measures;

^*Among couples that were debt concordant; underlined values indicate significant different by debt agreement, p<0.05

All couples were heterosexual and predominantly non-Latinx white (88%), with the remaining respondents Latinx (7%), and non-Latinx Black (5%). The majority of the respondents reported having either a high school degree or some college education (81%), and just over half (52%) were employed full-time. The average number of economic and material hardships experienced was 2.15 out of a total of 10 items with more than half of the sample reporting that they also experienced at least 2 or more material hardships during the past year.

Carrying credit card debt is common among these couples. More than half of the sample, (58%) reported non-zero debt with an average balance of $3,752. The outstanding debt balance between debt concordant and discordant couples was significantly different (p<0.05) - concordant couples reported lower balances compared to discordant couples. Twenty-five percent of the respondents indicated not having credit cards, and the remaining 17% reported no credit card debt because they did not carry a balance. More than half the couples (55%) were debt concordant. Reporting the same amounts of positive debt comprised the largest group at 41%, followed by 37% of couples who shared knowledge that the household held no credit cards. Approximately 22% of the concordant couples were aware that the household paid off their bill and held a zero balance. In addition, a majority of individuals reported using an income pooling system to manage their finances (61%), reported making joint decisions on big purchases (85%), and held joint bank accounts (82%).

Relationship satisfaction was high within our sample, averaging 8.32 out of a possible 10, with debt discordant couples reporting lower satisfaction. Women reported lower satisfaction levels than men and there was no statistical difference by concordance status, whereas men in debt concordant relationships reported higher averages levels of satisfaction. Average within couple differences in satisfaction varied significantly by concordance status with mean values under one point (0.873) for concordant couples and greater than one (1.379) for discordant couples.

Financial Management Characteristics, by Debt Concordance

Based on the debt concordance status of the couples, we then explored differences with regards to several relationship-specific practices often associated with household finances. Charts 1 and 2 display the summary statistics for the cross-tabulations of the management practices (Chart 1) or frequency of arguments (Chart 2) with the couples’ concordance status. Although more debt concordant couples reported using an income pooling system and making joint decisions, these differences were not statistically significantly different at conventional levels (p<0.05). Interestingly, debt discordant couples were more likely to have a joint bank account (85%) relative to debt concordant couples (80%). This may indicate that integrated financial management systems are not necessarily proxies for joint awareness of finances related to household debt.

Chart 1. Financial Management Measures, by Debt Concordance.

Note: Based on analytic data from MARS, n= 870; *** p<0.001, ** p<0.01, * p<0.05, + p<0.10, significant different by debt agreement.

Chart 2. Frequency of Financial Arguments, by Debt Concordance.

Note: Based on analytic data from MARS, n= 870

Debt concordant couples were also less likely to argue over managing money, bills, and debt relative to debt discordant couples; while 38% of debt concordant couples never argued over finances, 26% of debt discordant couples never argued over finances. Approximately 31% of debt discordant couples argued a few times a month or more over matters related to finances, while 21% of debt concordant couples argued over managing money a few times a month or more (χ²=18.29, p<0.000). Overall, apart from joint bank accounts which favored discordant couples, engagement in financial integration practices did not vary by concordance status. We also found that at least descriptively, debt concordant couples are more likely to compromise and argue less frequently about their finances relative to debt discordant couples.

Debt Concordance and Relationship Quality

Table 2 presents the estimates from the HLM models predicting relationship satisfaction. In Model 1, we present unconditional estimates of debt concordance on relationship satisfaction. In Model 2, we expand the model to include the current relationship status, as well as background demographic and socioeconomic factors. In Model 3, we added indicators of joint financial management practices. In the fourth column, Model 4, we adjusted for the frequency of arguments on financial issues.

Table 2.

HLM Models of Relationship Satisfaction on Consumer Debt Concordance and additional covariates

	(1)	(2)	(3)	(4)	(5)
Consumer debt concordance	0.382*	0.333*	0.310*	0.244+
	[0.152]	[0.144]	[0.141]	[0.139]
Current Relationship Status (ref. Current Cohab)
Directly Married		0.848**	0.651*	0.609*	0.632*
		[0.313]	[0.300]	[0.288]	[0.290]
Cohab, then Married		0.531+	0.367	0.433	0.443
		[0.302]	[0.279]	[0.273]	[0.272]
Compromise during arguments		0.962***	0.892***	0.784***	0.783***
		[0.140]	[0.134]	[0.134]	[0.135]
Financial Management Practices
Income Pooling System			0.16	0.161	0.16
			[0.123]	[0.119]	[0.119]
Joint decision on big purchases			0.666**	0.568*	0.557*
			[0.240]	[0.233]	[0.231]
Joint bank account			0.149	0.116	0.126
			[0.188]	[0.182]	[0.183]
Frequency of arguments over financial management (ref: Never)
A few times a year				−0.416***	−0.413***
				[0.118]	[0.119]
A few times a month or more				−0.972***	−0.977***
				[0.189]	[0.189]
Concordance on: (ref: Discordant)
Outstanding debt amount					0.324*
					[0.148]
Zero balance, paid off monthly					0.225
					[0.230]
No credit cards					0.036
					[0.246]
Fixed Intercept	8.114***	7.488***	6.971***	7.404***	7.431***
	[0.111]	[0.397]	[0.436]	[0.440]	[0.445]
Random Effects Parameters
Intercept	0.274**	0.177*	0.121	0.049	0.047
	[0.088]	[0.090]	[0.091]	[0.102]	[0.101]
Residual	0.212***	0.150**	0.159**	0.162**	0.162**
	[0.058]	[0.053]	[0.054]	[0.056]	[0.056]
F-stat	6.32	5.99	5.43	7.39	6.82
Observations	870	870	870	870	870
Number of groups	435	435	435	435	435

Note: Models 2, 3, 4, and 5 adjust for gender, race/ethnicity, educational attainment, employed status, logged household income, number of economic/material hardships, relationship length, relationship history, and total number of children in the household; robust standard errors in brackets;

^***p<0.001,

^**p<0.01,

^*p<0.05,

⁺p<0.10

In the basic unconditional model (Model 1), consumer debt concordance was associated with a 0.382-point increase in relationship satisfaction, which is equivalent to approximately 0.21 of a standard deviation change (.382/1.77). This estimate is also in line with descriptive statistics that found debt concordant couples reported higher relationship quality (debt concordant couples reported a mean score of 8.31 and debt discordant couples reported a mean score of 8.11). Including family background factors and relationship characteristics in Model 2 attenuated the consumer debt concordance estimate, which was now associated with a 0.33-point increase in relationship satisfaction (please see Appendix A for full model results). Another factor significantly associated with decreased relationship satisfaction was having an increased number of economic and material hardships, which was associated with a 0.12-point decrease in relationship satisfaction.

With the inclusion of relationship characteristics related to financial integration (Model 3), consumer debt concordance was still significantly associated with a 0.31-point increase in relationship satisfaction. In addition to the number of economic and material hardships, which was negatively associated with relationship satisfaction, directly marrying and making joint decisions on big purchases were both positively associated with relationship satisfaction. Compromising during arguments was also associated with a 0.892-point increase in relationship satisfaction.

With the inclusion of frequency of arguments in Model 4, debt concordance became weakly significantly associated with a 0.24-point increase in relationship satisfaction. These results indicate that arguing over finances explains some of the differential in relationship satisfaction by concordance status, given the concordance estimate was reduced by 7% from Model 3 to Model 4. We found that arguing over finances was significantly associated with decreased relationship satisfaction relative to couples who never argued over finances, as well as experiencing an increased number of economic and material hardships.

In the fifth column, we also present estimates from a model that replaced the dichotomous debt concordance measure with a disaggregated concordance measure in order to see if there was any variation in the types of concordance and relationship quality. The model estimates indicated that agreement on the amount of outstanding credit card debt was positively associated with overall relationship satisfaction (p<0.001) relative to debt discordant couples. Couples who were in agreement that the household held no credit cards or also in agreement on not carrying a balance were more likely to report better relationship quality - however, these estimates never reached conventional levels of significance. When debt discordance was interacted with current relationship status to examine whether there were different implications of relationship quality for debt concordant and debt discordant couples, there was no evidence that being in a cohabiting couple or having cohabited prior to marriage was differentially associated with lower relationship quality relative to marrying directly.

We are limited in our ability to examine the extent to which structural ambiguities exist in debt management; however, we posit that the range of quoted household debt values between partners may serve as an indirect measure. In separate analysis we replaced the continuous credit card debt measure with the difference in household debt (mean=0, std. error=235.8) reported between partners in Model 4 from Table 2. The magnitude of debt discordance was negatively associated with relationship satisfaction; however, the model estimate was not significant (B=−0.01, t-stat=−0.49), nor was the estimate on the debt concordance measure (B=0.06, t-stat=0.62) with its inclusion.

In additional analyses, two flexible measures of debt concordance were tested that changed the inclusivity of the concordance on the continuous credit card debt category. The first measure allowed couples to be concordant even if they did not agree on the outstanding amount, but both specified an amount within $500 of one another. The second most generous measure allowed for concordance if a couple reported values within $1000 of each other. When debt concordance was expanded to include couples who agree within $500, the total percentage increased to 60%, decreasing those in the discordance group to 40%, and when it included those within $1000 the concordance group grew to 63%. The results of these analyses, available in Appendix B, show similar patterns to those in Table 2. Given the strictest results were reported in Table 2, it is not surprising those model estimates are an upper bound. As the number of concordant couples grew, the magnitude of the statistically significant association between relationship quality and debt agreement either weakened or stayed the same.

Discussion and Conclusion

Given the increased significance and attention consumer debt has received among consumer and family scholars, we explored whether debt concordance among couples was associated with relationship quality among cohabiting and married couples with children. In a 2008 article, Pahl argues that consumer products like credit cards allow couples to maintain individual finances while also practicing collectivist money management systems. Credit card usage, Pahl claims, has led to the diminishing of pooled systems as the preferred way to manage household finances in order to maintain autonomy over one’s finances. While Pahl (2008) posits that credit cards have led to increased individualization of money management, the results of this study suggest that agreement related to the amount of household credit card debt enhances relationship satisfaction, above and beyond the financial management practices.

Overall, our model estimates suggest that debt concordance was associated with higher relationship quality. This relationship was linear and monotonic; however, we did not find that the association differed among those who reported quality below or above the median. We also found that economic characteristics of the couple, such as making joint decisions on purchases, the number of economic and financial hardships experienced, arguing over finances, and compromising within the relationship were associated with relationship quality. The disaggregated regression results also show that concordance on positive outstanding debt appears to be driving much of the relationship. This suggests that while outstanding debt amounts may be important for relationship functioning, mechanisms for communicating debt may be more important for improved relationship quality, especially among low-income couples with children. The fact that disagreements related to financial issues attenuate the debt concordance-relationship quality association suggests that couples may be debt concordant because they argue about their finances more frequently. When concordance is excluded from the model, the estimates on the measure of financial arguments remains statistically significant and increases.

This study extends previous research on consumer debt and the family by reframing the debt in the household discussion beyond whether it is the amount of outstanding debt influencing couples’ well-being (Dew 2007, 2008) to a focus on whether couple-level knowledge and agreement about debt and debt behaviors matter for relationship outcomes. This research also adds to research showing that distinguishing between socioeconomic resources and consumer products in household finance matters for intra-household family dynamics (Dew 2007), and that examining both partners in couples is useful for understanding household dynamics (Halpern-Meekin and Tach 2013). Three overarching themes that arise from this study include the high prevalence of debt concordance within the sample, the importance of objective economic measures to assess relationship quality, and that relationship characteristics (directly marrying, compromising, and having fewer arguments over finances) also predict relationship quality.

The first theme that emerged from this study was related to the proportion of debt concordant couples. Approximately 55% of couples in the study were debt concordant. That the majority of couples sampled were debt concordant not only indicates an awareness and knowledge of household debt, but also suggests that couples discussed their household debt. In addition to income pooling and joint decision-making, participation of both members in household financial management does not only include the positive or in-flow of money (e.g. savings and asset building), but also negative outflows of money. However, debt discordance was also high (45%). From a survey methods perspective, this signals that selecting one person to answer household financial questions may increase the collection of inaccurate financial data, increasing measurement error.

The second theme that emerged from this study is that both objective and subjective financial constructs can provide critical information related to relationship quality. Similar to previous work, we found strong evidence that self-assessed and more subjective measures related to the relationship such as how arguments were handled, the frequency of arguments, and making joint decisions on big purchases were associated with relationship satisfaction. However, we also found that in addition to agreement on the amount of outstanding debt, the magnitude of economic and material hardships of the couples matters. Among this sample of relatively low-income couples with children, we found no evidence that income itself was significantly associated with relationship quality. These findings highlight the need to incorporate several different metrics of economic characteristics in assessing couple-level wellbeing.

The third theme that emerged from this study was that relationship characteristics, such as directly marrying, compromising within the relationship, and arguing over finances, were significantly associated with relationship quality. Relative to cohabiting couples, couples who directly married were more likely to report higher relationship satisfaction. In addition, couples who compromised were more likely to have increased relationship satisfaction. On the other hand, couples who argued over their finances were significantly more likely to report lower relationship satisfaction. The inclusion of this measure attenuated the debt concordance-relationship quality association. This result suggests that one reason couples may be debt concordant is because they argue about their finances more frequently, are more aware of the household balance, or are monitoring it more closely than couples who do not argue as often or at all.

Our study is not without limitations. First, although our data and analysis allow us to understand the characteristics of couples who are debt concordant, we are unable to explain why these couples were more likely to agree. In further sensitivity analysis debt concordance was regressed on the full model of covariates including a measure of within couple difference in relationship status. There is suggestive evidence that frequency of finance-related arguments was associated with concordance; this suggests that intra-couple communication, even argumentative discussions, may be critical to transparency about household debt. In addition, the survey instructions indicated that the survey was to be completed alone and provided respondents with separate logins and passwords, there is no way to verify whether couples collaborated on answering questions on finances, including the outstanding debt amount. If some communication did occur, it may be possible that the results are an upper bound and biased towards couples with better relationship quality, especially if colluders were also likely to report fewer arguments and greater satisfaction. The survey was also fielded in 2006, less than a year prior to the housing bust and the start of the Great Recession (2007–2009). As credit markets contracted, many low-income households lost access to a potential safety net during a period of high unemployment and lost wages.

Throughout the recession and for several years after, it was extremely difficult to obtain or increase credit lines. Our findings should therefore be interpreted within the context of the booming economy pre-recession. In addition, study conclusions are only limited to heterosexual couples and there is overrepresentation non-Latinx white couples in the sample. Therefore, we were limited in our ability to conduct disaggregated analyses by race and ethnicity. Finally, an additional limitation is that the study was based on cross-sectional data. Therefore, we do not make any causal claims regarding the direction of causality or the strong positive correlation between debt concordance and relationship satisfaction.

In conclusion, this study highlights several important areas for future research. First, empirical research on household finances of a couple really benefit from surveying both members of the partnership, as other recent studies have shown (Halpern-Meekin and Tach 2013). Second, debt and debt behavior may have implications for relationship outcomes far beyond the household balance sheet. Finally, financial measures can serve as objective proxies of household behavior. In particular, they may signal the current state of affairs that may be hard to capture from subjective measures of relationship expectations and satisfaction. Growing levels of household debt have been increasingly tied to societal levels of economic inequality. Low-income households with children have not been immune to this changing financial landscape (Bird, Hagstrom, and Wild 1999; Pressman and Scott 2009). What couples know about their current state of their finances, in addition to how couples choose to manage these finances, can be reflected in their relationship quality.

Appendix A.

HLM Models of Relationship Satisfaction on Consumer Debt Concordance and additional covariates

	(1)	(2)	(3)	(4)	(5)
Consumer debt concordance	0.382*	0.333*	0.310*	0.244+
	[0.152]	[0.144]	[0.141]	[0.139]
Female		−0.193+	−0.214*	−0.178+	−0.175+
		[0.103]	[0.104]	[0.102]	[0.102]
Race/Ethnicity (ref: Non-Latinx White)
Non-Latinx Black		−0.235	−0.196	−0.090	−0.088
		[0.373]	[0.354]	[0.345]	[0.345]
Latinx		−0.097	−0.076	−0.030	−0.044
		[0.218]	[0.219]	[0.219]	[0.220]
Educational Attainment (ref: Less than high school)
High School		−0.088	−0.104	−0.100	−0.090
		[0.206]	[0.207]	[0.204]	[0.204]
Some College		−0.212	−0.243	−0.244	−0.233
		[0.213]	[0.216]	[0.213]	[0.212]
Bachelors or more		−0.337	−0.379	−0.439+	−0.404
		[0.257]	[0.261]	[0.258]	[0.254]
Currently employed full-time		0.173	0.153	0.183	0.188
		[0.127]	[0.127]	[0.127]	[0.128]
Household Income (ln)		0.008	0.008	0.014	0.014
		[0.017]	[0.017]	[0.017]	[0.017]
Credit card balance (ln)		−0.003	−0.002	0.009	0.000
		[0.017]	[0.017]	[0.017]	[0.020]
Number of economic/material hardships		−0.124***	−0.114***	−0.087**	−0.089**
		[0.031]	[0.031]	[0.031]	[0.031]
Current Relationship Status (ref. Current Cohab)
Directly Married		0.848**	0.651*	0.609*	0.632*
		[0.313]	[0.300]	[0.288]	[0.290]
Cohab, then Married		0.531+	0.367	0.433	0.443
		[0.302]	[0.279]	[0.273]	[0.272]
Total Number of Children in Household		−0.048	−0.056	−0.048	−0.051
		[0.073]	[0.071]	[0.067]	[0.067]
Relationship Length		−0.001	−0.001	−0.001	−0.001
		[0.001]	[0.001]	[0.001]	[0.001]
Previously Married		0.098	0.047	0.008	0.014
		[0.138]	[0.136]	[0.134]	[0.134]
Compromise during arguments		0.962***	0.892***	0.784***	0.783***
		[0.140]	[0.134]	[0.134]	[0.135]
Financial Management Practices
Income Pooling System			0.160	0.161	0.160
			[0.123]	[0.119]	[0.119]
Joint decision on big purchases			0.666**	0.568*	0.557*
			[0.240]	[0.233]	[0.231]
Joint bank account			0.149	0.116	0.126
			[0.188]	[0.182]	[0.183]
Frequency of arguments over financial management (ref: Never)
A few times a year				−0.416***	−0.413***
				[0.118]	[0.119]
A few times a month or more				−0.972***	−0.977***
				[0.189]	[0.189]
Concordance on: (ref: Discordant)
Outstanding debt amount					0.324*
					[0.148]
Zero balance, paid off monthly					0.225
					[0.230]
No credit cards					0.036
					[0.246]
Constant	8.114***	7.488***	6.971***	7.404***	7.431***
	[0.111]	[0.397]	[0.436]	[0.440]	[0.445]
Intercept	0.274**	0.177*	0.121	0.049	0.047
	[0.088]	[0.090]	[0.091]	[0.102]	[0.101]
Residual	0.212***	0.150**	0.159**	0.162**	0.162**
	[0.058]	[0.053]	[0.054]	[0.056]	[0.056]
Observations	870	870	870	870	870
Number of groups	435	435	435	435	435

Robust standard errors in brackets

^***p<0.001,

^**p<0.01,

^*p<0.05,

⁺p<0.10

Appendix B.

HLM Models of Relationship Satisfaction on Consumer Debt Concordance and additional covariates, alternative concordance measures

		Concordance within $500					Concordance within $1000
Models	(1)	(2)	(3)	(4)	(5)	(1)	(2)	(3)	(4)	(5)
Consumer debt concordance	0.345*	0.323*	0.303*	0.234+		0.382*	0.363*	0.343*	0.267+
	[0.156]	[0.145]	[0.142]	[0.140]		[0.160]	[0.146]	[0.142]	[0.142]
Female		−0.194+	−0.215*	−0.179+	−0.177+		−0.193+	−0.214*	−0.179+	−0.176+
		[0.103]	[0.104]	[0.103]	[0.102]		[0.103]	[0.104]	[0.103]	[0.102]
Race/Ethnicity (ref: Non-Latinx White)
Non-Latinx Black		−0.237	−0.200	−0.092	−0.090		−0.235	−0.199	−0.092	−0.090
		[0.373]	[0.354]	[0.345]	[0.345]		[0.372]	[0.353]	[0.344]	[0.344]
Latinx		−0.092	−0.072	−0.027	−0.037		−0.084	−0.064	−0.022	−0.032
		[0.218]	[0.220]	[0.219]	[0.220]		[0.218]	[0.219]	[0.218]	[0.220]
Educational Attainment (ref: Less than high school)
High School		−0.089	−0.105	−0.101	−0.092		−0.091	−0.107	−0.103	−0.094
		[0.206]	[0.207]	[0.205]	[0.205]		[0.206]	[0.207]	[0.204]	[0.205]
Some College		−0.213	−0.244	−0.244	−0.235		−0.217	−0.248	−0.247	−0.238
		[0.214]	[0.216]	[0.213]	[0.213]		[0.213]	[0.216]	[0.213]	[0.212]
Bachelors or more		−0.324	−0.366	−0.429+	−0.395		−0.324	−0.365	−0.428+	−0.392
		[0.258]	[0.262]	[0.259]	[0.254]		[0.258]	[0.262]	[0.259]	[0.254]
Currently employed full-time		0.175	0.154	0.184	0.188		0.176	0.155	0.185	0.190
		[0.128]	[0.128]	[0.128]	[0.128]		[0.128]	[0.128]	[0.128]	[0.128]
Household Income (ln)		0.007	0.007	0.013	0.013		0.007	0.007	0.013	0.013
		[0.017]	[0.017]	[0.017]	[0.017]		[0.017]	[0.017]	[0.017]	[0.017]
Credit card balance (ln)		−0.005	−0.004	0.007	0.001		−0.006	−0.005	0.006	−0.002
		[0.017]	[0.017]	[0.016]	[0.020]		[0.016]	[0.017]	[0.016]	[0.020]
Number of economic/material hardships		−0.125***	−0.115***	−0.088**	−0.091**		−0.125***	−0.115***	−0.088**	−0.091**
		[0.031]	[0.031]	[0.031]	[0.031]		[0.031]	[0.031]	[0.031]	[0.031]
Current Relationship Status (ref. Current Cohab)
Directly Married		0.834**	0.640*	0.600*	0.619*		0.825**	0.637*	0.598*	0.616*
		[0.312]	[0.299]	[0.288]	[0.289]		[0.311]	[0.299]	[0.287]	[0.289]
Cohab, then Married		0.518+	0.357	0.424	0.431		0.509+	0.353	0.421	0.427
		[0.300]	[0.277]	[0.272]	[0.271]		[0.299]	[0.276]	[0.271]	[0.270]
Total Number of Children in Household		−0.043	−0.051	−0.044	−0.046		−0.043	−0.051	−0.044	−0.046
		[0.073]	[0.070]	[0.067]	[0.067]		[0.073]	[0.070]	[0.067]	[0.067]
Relationship Length		−0.001	−0.001	−0.001	−0.001		−0.001	−0.001	−0.001	−0.001
		[0.001]	[0.001]	[0.001]	[0.001]		[0.001]	[0.001]	[0.001]	[0.001]
Previously Married		0.100	0.049	0.009	0.012		0.102	0.051	0.011	0.016
		[0.138]	[0.136]	[0.134]	[0.134]		[0.138]	[0.136]	[0.134]	[0.134]
Compromise during arguments		0.968***	0.897***	0.788***	0.789***		0.969***	0.898***	0.789***	0.790***
		[0.140]	[0.134]	[0.135]	[0.135]		[0.140]	[0.134]	[0.134]	[0.135]
Financial Management Practices
Income Pooling System			0.160	0.161	0.160			0.159	0.160	0.158
			[0.123]	[0.119]	[0.119]			[0.123]	[0.119]	[0.119]
Joint decision on big purchases			0.668**	0.571*	0.563*			0.671**	0.574*	0.567*
			[0.240]	[0.232]	[0.231]			[0.239]	[0.232]	[0.230]
Joint bank account			0.144	0.113	0.125			0.135	0.106	0.116
			[0.188]	[0.182]	[0.183]			[0.188]	[0.182]	[0.182]
Frequency of arguments over financial management (ref: Never)
A few times a year				−0.418***	−0.418***				−0.419***	−0.418***
				[0.118]	[0.118]				[0.118]	[0.118]
A few times a month or more				−0.971***	−0.976***				−0.966***	−0.970***
				[0.190]	[0.190]				[0.190]	[0.190]
Concordance on: (ref: Discordant)
Outstanding debt amount					0.288*					0.329*
					[0.144]					[0.147]
Zero balance, paid off monthly					0.237					0.254
					[0.235]					[0.235]
No credit cards					0.042					0.061
					[0.248]					[0.249]
Constant	8.116***	7.491***	6.972***	7.409***	7.426***	8.083***	7.460***	6.940***	7.382***	7.406***
	[0.122]	[0.396]	[0.437]	[0.442]	[0.447]	[0.130]	[0.396]	[0.437]	[0.443]	[0.448]
Intercept
	0.277**	0.177*	0.121	0.049	0.047	0.275**	0.175+	0.119	0.048	0.046
Residual	[0.087]	[0.089]	[0.091]	[0.101]	[0.101]	[0.087]	[0.090]	[0.092]	[0.101]	[0.101]
	0.212***	0.150**	0.159**	0.162**	0.162**	0.212***	0.150**	0.159**	0.162**	0.162**
	[0.058]	[0.053]	[0.054]	[0.056]	[0.056]	[0.058]	[0.053]	[0.054]	[0.056]	[0.056]
Observations	870	870	870	870	870	870	870	870	870	870
Number of groups	435	435	435	435	435	435	435	435	435	435

Robust standard errors in brackets

^***p<0.001,

^**p<0.01,

^*p<0.05,

⁺p<0.10