Different measures of money atttitudes have been developed through out the years. Yamauchi and Templer (1982) developed the first well-known money attitudes scale (MAS). Other important measures of money attitudes include Furnham’s (1984) Money Beliefs and Behavior Scale and Tang’s (1992) money ethic scale. Although with somewhat different factorial structures, which have evolved as more research with these and other attitude measures was done (e.g., Tang, 1993, 1995; Roberts and Sepulveda, 1999; Klontz et al., 2011; Lay and Furnham, 2019), the most important dimensions that emerged from the different attitudes scales can be described in terms of the cognitive, affective, and behavioral components of the tripartite model of attitudes first proposed by Rosenberg and Hovland (1960; see also, Eagly and Chaiken, 1993; Chatterjee et al., 2018). For instance, the main dimensions of the MAS include (a) the perception of money as a source of freedom/power and achievement (cognitive component); (b) The association of money with financial planning and budget (behavior component); and (c) money as a source of distrust, suspicious and anxiety as well as protection from anxiety (affective component).

Since one of our main goals is to evaluate how money attitudes impact financial behaviors, the measure used herein focus on the behavioral component of these attitudes (i.e., how much consumers take care of and monitor their money) as the factor that may influence one or more of the three categories of financial behavior investigated. Indeed, although Rosenberg and Hovland (1960; see also Kaiser and Wilson, 2019) argued that the three components of attitudes correspond to manifestations of the same latent variable (i.e., attitudes toward money), the attitude’s behavioral component is likely to be a more direct indicator of related financial behaviors than the cognitive and affective components.

Prior research on money attitudes suggests individuals demonstrate a variety of predisposed responses toward money. Money has different meanings and serves different purposes for different individuals, leading them to act differently toward it. Importantly for this research, attitudes regarding one’s actions toward money vary considerably as well, with some demonstrating ease in spending and accruing debt and others seeming more anxious and more devoted to saving money.

Money attitudes have been shown to have a significant positive impact on financial management behavior among young adults (Qamar et al., 2016). In their study, Qamar et al. used a money attitude questionnaire that included measures of money avoidance, worship, status and vigilance (Klontz et al., 2011) and found that 20.9% of the personal financial management behavior was explained by money attitudes.

Dowling et al. (2009) showed that financial management practices and money attitudes – measured in terms of (a) the importance ascribed to the ownership of material goods, (b) the extent to which one uses money as a standard of comparison with others; and (c) the extent to which individuals think and worry about money – predicted financial problems among young male Australian workers.

Similarly, Klontz and Britt (2012) showed that core attitudes and beliefs about money drive financial behaviors (see also Kahler and Fox, 2005). In their study three belief patterns (money avoidance, money status, and money worship) were not only associated with lower levels of total assets, lower earnings, and higher measures of revolving credit but could also foresee disordered money practices such as impulsive purchasing and financial dependence. In contrast, attitudes and convictions favoring money vigilance, including frugality, caution, and anxiety about money, appeared to be protective against poor finances and dangerous financial practices.

More recently, Sabri et al. (2020), studied the determinants of employees’ financial well-being in Malaysia and found out that money attitudes, substantially contributed to employee’s fulfilment of current and ongoing financial obligations (above and beyond financial practices, self-efficacy, and emotional coping).

Money attitudes have also been shown to be a predictor of credit card debt. Specifically, the money attitude of effort/ability appears to play an important role (together with attitude toward credit) in distinguishing between more indebted students (with four or more credit cards) and less indebted ones (with one to three credit cards; Hayhoe et al., 1999). Likewise, good financial management practices (e.g., budgeting, saving, and regulating spending; see Godwin and Koonce, 1992) were also shown to be a main predictor of debt levels (Lea et al., 1995; Donnelly et al., 2012; Ksendzova et al., 2017).

Taken together, these studies provide strong evidence concerning the impact of money attitudes on consumers’ financial behavior and show that both attitudes and financial behaviors are important predictors of indebtedness. However, in contrast with the present study, most extant research has not distinguished between overindebted and non-overindebted households and has not considered the possible differences (in terms of money attitudes, financial behaviors and their relation) between households who became overindebted due to internal (individualistic) versus external (situational) factors.

In the study here reported, we try to shed some light on these issues by assessing attitudes toward money of overindebted and non-overindebted consumers through its behavioral component (i.e., how much consumers take care and keep track of their money). Furthermore, consumers’ financial management behavior was assessed by means of a cash flow management scale (Godwin and Koonce, 1992) that includes three different categories of behaviors: record keeping, adjusting balance, and monitoring balance.

Our main goals are to evaluate (a) how money attitudes and money management behaviors differ across non-overindebted and (two types of) overindebted participants; and (b) how our measure of money attitudes is related to consumers’ money management behaviors. In short, we aim to address whether overindebted and non-overindebted individuals differ in a behavioral component of money attitudes, and whether this impacts financial management behavior above and beyond consumers’ indebtedness status.

Our sample consisted of 365 overindebted and non-overindebted participants. Overindebted participants were consumers who sought assistance and counseling with DECO (a Portuguese NGO for consumer defense) throughout 2017 regarding their problem with overindebtedness. Non-overindebted participants were collected through convenience sampling and asked if they were in a situation of overindebtedness, in which case they were recoded as overindebted (six participants were recoded this way, amounting to 236 overindebted and 129 non-overindebted consumers in the final sample). Fifty-eight participants identified as male and 76 as female, the remainder of the sample did not answer. Participants’ mean age was 50 years (SD = 15.58), the majority (73.94%) did not have a degree and lived alone (30.58%) or with another person (34.17%). Their household mean income per capita (i.e., total income divided by number of people in the household) was 796,93€ (SD = 654.67). Additional information is presented on Table 1.

Overindebted participants responded to the questionnaire in either a paper and pencil format (those who went to DECO for a consultation) or in an editable computer file sent to them by mail (those who contacted DECO through their website or by email). Non-overindebted participants responded to the questionnaire in a paper and pencil format.

Participants answered socio-demographic questions (marital status, level of schooling, professional status, number of people in the household) and questions concerning economic aspects of their life (monthly income, monthly expenses, and credit product installments). Reported causes of overindebtedness were collected by DECO and for some participants we were able to pair this information with our data (using codes that allowed to guarantee consumers anonymity). Participants identified one or more causes for their overindebtedness situation from several pre-defined options. These causes were categorized into two groups: ‘‘Internal causes’’ and ‘‘External causes.’’ The first included participants that identified ‘‘poor money management’’ and ‘‘excessive use of credit’’ as causes. The second included participants that identified ‘‘salary cuts,’’ ‘‘unemployment (self or spouse),’’ ‘‘divorce,’’ ‘‘death in the family,’’ and ‘‘birth in the family’’ (among other external causes), as causes for their overindebtedness. From the original 236 overindebted participants, 95 reported external causes and 36 reported internal causes for their overindebtedness. For the remaining 105 participants we could not identify the overindebtedness causes because participants did not report them or because we were not given access to this information due to confidentiality issues¹.

Money attitudes were measured by asking participants to “Please indicate your degree of agreement with the following affirmations” referring to two items originally used by Lea et al. (1993), and obtaining the mean from both responses. The items used were “I am careless with money,” “I find it hard to keep track of my money” (Cronbach’s alpha 0.85). Participants responded to these items using a 5-point rating scale, from 1 (Very strongly agree) to 5 (Very strongly disagree). Money attitude was the mean between both responses. This measure was aimed at capturing the behavioral component of the multifaceted concept of money attitudes, which includes, in addition, a cognitive and an affective component (according to the tripartite model; Rosenberg and Hovland, 1960; Chatterjee et al., 2018).

Financial management behaviors were measured with 11 items taken from the Cash Flow Management scale described in Godwin and Koonce (1992) to which participants responded, using a 5-point scale from 1 (Never) to 5 (Always), to the question “Please indicate how often do you:” followed by the 11 items. A principal components analysis using the full sample yielded three components. The first component, with an eigenvalue of 5.56 and a Cronbach alpha of 0.89, was named “Record keeping” and includes the following financial behaviors: “Recording in writing most spending,” “Recording in writing your actual income,” “Assessing the amount of money spent on fixed expenses (rent, car, and payments, etc.),” “Assessing total expenses,” and “Assessing the amount of money spent on flexible expenses (food, clothing, and recreation, etc.).” The second component, with an eigenvalue of 1.26 and a Cronbach alpha of 0.70, was named “Adjusting balance,” and includes the following financial behaviors: “Thinking of ways to increase your income to match your needs” and “Try to think of ways to decrease your expenses to match your income.” The third component, with an eigenvalue of 1.02, and a Cronbach alpha of 0.78, was named “Monitoring balance” and includes the following behaviors: “Monitoring your spending to see if it is within your income,” “Monitoring your spending to see if it is in line with what you expected,”, Assessing the amount of money you can use during an emergency” and “Assessing the value of things you own.” Components values were the mean responses to their respective items. Money attitudes were measured prior to financial management behaviors. Finally, the questionnaire also included other measures collected for different research purposes, namely on the perceived causes of over-indebtedness causes, sleep quality, perceived health, feelings and locus of control, life satisfaction, financial satisfaction, well-being, and attitudes toward poverty.

A one-way analysis of covariance (ANCOVA) was performed with three levels of indebtedness status (overindebted as the result of internal causes; overindebted as the result of external causes; non-overindebted) as a between-participants factor and household income per capita and level of schooling as covariates. A main effect of indebtedness status, F(1, 198) = 16.54, p < 0.001, ηp2 = 0.14, showed that non-overindebted participants have more positive attitudes toward money (i.e., a disposition to care and monitor their money; M = 4.29; SE = 0.29) than participants overindebted by external causes (M = 3.87; SE = 0.14), F(1, 198) = 2.38, p < 0.001, ηp2 = 0.01, which in turn showed more positive attitudes than overindebted participants due to internal causes (M = 2.53, SE = 0.30), F(1, 198) = 1.92, p < 0.001, ηp2 = 0.01. No other effects were significant. A similar result is found when the analysis is performed without the covariates, F(1, 203) = 18.32, p < 0.001, ηp2 = 0.15.

In sum, non-overindebted consumers appear to show a positive attitude (as indicated by a mean response above the scale mid-point), consumers who become overindebted due to external factors showed weaker money attitudes (mean response close to the mid-point of the scale) and consumers who became overindebted due to internal factors showed the weakest money attitudes (below the mid-point of the scale).

We performed a 3 indebtedness status (overindebted as the result of internal causes; overindebted as the result of external causes; non-overindebted) × 3 financial management behaviors (record keeping, adjusting balance, and monitoring balance) mixed measures ANCOVA with the first factor between-participants, the second within participants and household income per capita and level of schooling as covariates.

The analysis yielded a main effect of financial behaviors, F(2, 400) = 14.26, p < 0.001, ηp2 = 0.07. Adjusting balance was the more frequently reported behavior (M = 3.83, SE = 0.18), followed by monitoring balance (M = 3.16, SE = 0.18), and record keeping (M = 3.08, SE = 0.18). Post hoc comparisons (with a Bonferroni correction) revealed that while levels of monitoring balance and record keeping were not significantly different, both were significantly lower than adjusting balance (ps < 0.001).

There was also a significant interaction between overindebtedness status and financial management behaviors, F(4, 400) = 10.44, p < 0.001, ηp2 = 0.09. Planned comparisons showed that this effect was driven by significant differences in monitoring balance, F(2, 200) = 12.29, p < 0.001, ηp2 = 0.11, with non-overindebted participants reporting higher levels of monitoring balance (M = 3.81, SE = 0.09) than participants overindebted due to internal (M = 2.88, SE = 0.21), F(1, 200) = 2.02, p < 0.001, ηp2 = 0.01 and external causes (M = 3.21, SE = 0.12), F(1, 200) = 1.94, p < 0.001, ηp2 = 0.01. No other comparisons between conditions reached significance (all ps > 0.05).

Finally, there was a significant interaction between household income per capita and financial management behaviors, F(2, 400) = 3.25, p = 0.040, ηp2 = 0.02, such that higher household income predicts less financial behaviors of adjusting balance (β = −0.145, p = 0.050). The same analysis without the covariates yields a similar result, with a significant main effect of financial behaviors, F(2, 430) = 56.03, p < 0.001, ηp2 = 0.21, and a significant interaction between financial behaviors and overindebtedness status, F(4, 430) = p < 0.001, ηp2 = 0.13 (Figure 1).

In sum, although the frequency with which consumers adopted different kinds of financial management behaviors varies, we found no differences in the reported frequency of financial behaviors between consumers overindebted due to external and internal causes, and non-overindebted consumers, except for the monitoring balance behaviors. Overindebted consumers, regardless of the reported causes, engage significantly less in these kinds of behaviors than non-overindebted consumers. Therefore, monitoring balance behaviors (i.e., continuous monitoring of actual and future expenses in relation to one’s own income or wealth) appears to be associated with keeping expenses and debt service within manageable levels.

In line with previous analyses, zero-order correlations (see Table 2) show that indebtedness status (not being overindebted was coded 0 and being overindebted was coded 1) is associated with weaker money attitudes and poor monitoring balance. Naturally, overindebted consumers are also associated with larger debt-to-income ratios. Money attitudes are positively correlated with all measures of financial behaviors. In other words, higher levels of money attitudes are associated with better money management. The three measures of financial management behaviors are positively correlated with each other. Finally, money attitudes are negatively associated to debt-to-income ratio.

To further explore the relationship between our measure of the behavioral component of money attitudes and financial management behaviors, a multiple regression analysis was conducted to test the predictive role of overindebtedness and money attitudes (i.e., a disposition to care and monitor their money) on record keeping, adjusting balance, and monitoring balance behaviors, controlling for household income per capita and education level, using MPlus 7.2 (Muthén and Muthén, 1998–2012). Based on the results of the previous analyses, significant correlations between predictors and between criterion variables were included in the model. To evaluate model fit, the following indexes and criteria were used: the comparative fit index (CFI) above 0.95, the root mean square error of approximation (RMSEA), and the standardized root mean residual (SRMR) below 0.08, all indicative of a good fit (Hu and Bentler, 1999; Kline, 2011).

The multiple regression model fit the data well: χ²(2) = 2.03, p = 0.36; CFI = 1.00; RMSEA = 0.01, 90% CI: 0.00, 0.10; SRMR = 0.02. As depicted in Figure 2, results revealed that overindebtedness and money attitudes are negatively associated with each other, but both significantly predicted all three financial management behaviors analyzed. More specifically, both overindebtedness and money attitudes were associated with higher levels of record keeping (B = 0.30, p = 0.023 / β = 0.13, p = 0.022 and B = 0.37, p < 0.001 / β = 0.39, p < 0.001, respectively), and adjusting balance (B = 0.26, p = 0.019 / β = 0.14, p = 0.018 and B = 0.29, p < 0.001 / β = 0.36, p < 0.001, respectively). However, while money attitudes were also associated with higher levels of monitoring balance (B = 0.43, p < 0.001 / β = 0.50, p < 0.001), overindebtedness was associated with lower levels of this financial management behavior (B = −32, p = 0.003 / β = −0.16, p = 0.002). Regarding the covariates, household income per capita was associated with lower levels of adjusting balance behavior, while education level did not predict any of the financial management behaviors. The whole model, respectively, explained 15.5, 16.2, and 31.5% of the variance of record keeping, adjusting balance, and monitoring balance.

Next, to account for the different causes of overindebtedness, we once more divided the sample in three groups – (1) overindebted due to internal causes, (2) overindebted due to external causes, and (3) non-overindebted participants – and performed a multiple group analysis with AMOS (v. 26) to test if the model significantly differed between the three groups. An unconstrained multiple group model, with all paths allowed to be freely estimated across the three groups was compared to a model where all paths were constrained to be equivalent across groups. Results of this analysis revealed a non-significant chi-square difference between the unconstrained and the constrained models: Δχ²(18) = 23.36, p = 0.177, indicating that the model does not vary significantly between the three groups². Descriptive statistics (means and standard deviations) of money attitudes and financial management behaviors by group are presented in Table 1 of the Supplementary Materials.

In addition, despite the high percentage of missing data on participants’ gender (63.3%), we analyzed the potential moderating role of gender in the proposed model, by performing another multi-group analysis with the sample for which this data was available, using AMOS (v. 26). An unconstrained multi-group model, with all paths allowed to be freely estimated for both men and women was compared to a model where all paths were constrained to be equivalent for both groups. Results of this analysis also showed a non-significant chi-square difference between the unconstrained and the constrained models: Δχ²(9) = 7.86, p = 0.549, indicating that the model does not vary significantly between men and women. Descriptive statistics (means and standard deviations) of money attitudes and financial management behaviors by gender are presented in Table 2 of the Supplementary Materials.

At last, we also analyzed the potential moderating role of participants’ age in associations between money attitudes and the three financial management behaviors, using the PROCESS macro (v. 3) for SPSS Model 1 (Hayes, 2018), controlling for indebtedness status, household income per capita, and education level. Results did not reveal a significant interaction between participants’ age and money attitudes. Thus, age was not a moderator of the hypothesized associations.

Money attitudes have a multifaceted nature with different dimensions (e.g., Yamauchi and Templer, 1982), which according to the tripartite model of attitudes (Rosenberg and Hovland, 1960; Chatterjee et al., 2018), may be classified in three components (or indicators): a cognitive (e.g., money as achievement), an affective (money as a source of anxiety), and a behavioral component (budget and money monitoring). However, research using different measures of money attitudes, such as Yamauchi and Templer’s (1982) MAS, and Furnham’s (1984) Money Beliefs and Behavior Scale, often ignore these instruments multifactorial dimensions and use a money attitude index across all (or part of the) scales items (e.g., Qamar et al., 2016; Sabri et al., 2020). In our case, the money attitude measure only tapped the behavioral component of this attitude and included only two items. Although single-item measures have been used before to assess similar concepts (e.g., Almenberg et al., 2018 measured attitudes toward debt by asking one simple question “do you feel uncomfortable with having debt?”), these are important limitations of the present study that should be addressed in future research by using multidimensional measures to assess the three components of the money attitudes with multiple items per component. We speculate that the affective component of money attitudes, in particular, is likely to play an important role given that overindebtedness is often associated with depression, anxiety, and social stigma.

The correlational nature of the study reported herein, does not make it possible to ascertain causality. The causality issue is particularly troublesome for the regression analyses conducted to explore the relationship between our measure of money attitudes and financial management behaviors. Even so, given the more abstract nature of the attitude items and the more specific and concrete nature of the financial behavior items, having the former predicting the latter is in line with a considerable amount of literature on attitude-behavior consistency, according to which general predispositions or attitudes are used to predict behavior (e.g., Maio and Haddock, 2004; Fishbein and Ajzen, 2010). In other words, it appears to be theoretically sounder to consider and test the hypothesis that money attitudes predict engagement in specific money management behaviors, than the inverse: a specific financial behavior being the cause of a more general money-related attitude. Nevertheless, we acknowledge that it is not possible in the present study to fully disentangle the two causal directions. According to the self-determination theory (e.g., Bem, 1972), it might be the case that at least some consumers used their own financial behavior to infer their dispositions toward money (even if the money attitude measure was assessed before the financial management behaviors in our questionnaire). Regardless of which causal direction is dominant or stronger, the association between money attitudes and the three types of financial management behaviors suggests that interventions to improve one of them could have beneficial effects on the other.

Future research could use longitudinal designs (with at least two data collection waves) to measure the impact of debt-related attitudes on money management behavioral tasks, to more clearly ascertain causality.

Given the self-reported nature of our measures, response social desirability may have biased our results to a certain extent. Future research should thus replicate our findings using a paradigm more robust to social desirability bias, which can achieve considerable lengths in surveys on sensitive topics (Tourangeau and Yan, 2007; Gittelman et al., 2015). This could be achieved via indirect or implicit indicators of attitudes and behaviors or more easily by including a scale to account for individual differences in social desirability responding, so that its effect can be statistically controlled for.

Finally, although we controlled for differences in level of education and income (and tested for the impact of age and gender in the regression model) between overindebted and non-overindebted consumers in our sample, it is always possible that the two groups differ in yet other variables (besides the indebtedness status). These other differences may have contributed to explaining the reported findings. In the same vein, as participants completed other questionnaires before responding to the measures used in the present paper, their influence on these measures cannot be excluded.