Financial knowledge, financial confidence and learning capacity on financial behavior: a Canadian study

Abstract This study examines the relationships between financial knowledge, confidence, learning capacity, education and other sociodemographic information and financial behavior. A structural equation model is used to analyze the relationships between the study variables and to obtain a more comprehensive understanding of the factors linked to poor financial behavior among a large Canadian sample. The main findings showed that financial confidence plays a crucial role in explaining financial behavior and that learning capacity explains financial confidence. Overall, our results suggest that financial education should be considerably improved and that additional focus should be placed on financial confidence and individual’s learning capacity in order to mitigate existing financial difficulties, prevent new problems from arising, and develop and implement constructive strategies to achieve specific financial goals. This study contributes to previous research on financial literacy by demonstrating the influence of learning capacity on financial confidence and financial behavior.

manage one's financial resources effectively for lifetime financial security" (Hastings et al., 2013;Remund, 2010).. However, Oehler et al. (2018, p. 206) have argued that financial literacy is "not only the knowledge and understanding of financial concepts but also encompasses the skills, motivation and confidence to apply such knowledge in order to make effective decisions." While financial literacy is assumed to encourage desirable financial behavior (Potrich & Vieira, 2018) and improve financial well-being , many individuals appear to lack the financial literacy required to make appropriate and cost-effective financial decisions (Fisch et al., 2019;Lusardi & Mitchell, 2011b;Mandell & Klein, 2009). Researchers have found that financial illiteracy is a widespread international phenomenon, often characterized by poor financial behavior such as low or negative rates of personal savings, failing to invest wisely, and running out of money-and these missteps may lead to serious and often irreversible financial mistakes (Gathergood & Weber, 2017;Kim et al., 2019Kim et al., , 2020Lusardi & Tufano, 2015) While previous research findings are almost unanimous on the lack of financial literacy among the global population and on the need to improve financial behavior, the causes of financial illiteracy and the intervention strategies required to remedy it are much less explicit. While some studies found that financial knowledge leads to better financial behavior (Allgood & Walstad, 2016;Mountain et al., 2020), other researchers, such as Mandell and Klein (2009), found that students who took courses in personal financial management did not exhibit better financial behavior in subsequent years than those who did not take these courses. In this regard, several authors (Shim et al., 2009;Taft et al., 2013;Xiao et al., 2011) have argued the importance of exploring the possibility that psychological and cognitive factors might have positive effects on financial behavior. One of the psychological elements put forward to explain financial behavior is self-confidence, while learning capacity represents a cognitive factor of interest (Mudzingiri et al., 2018). Several studies also addressed sociodemographic variables such as age, gender, ethnicity and level of education. However, these have not been integrated into a single model to understand how their interrelationships may affect financial behavior. Furthermore, to our knowledge, no previous study has verified how learning capacity could act on financial behavior and could affect financial confidence. Therefore, the objective of this study is to examine the impact of financial knowledge, confidence, learning capacity and socio-demographics on the financial behaviors of university students, through the use of a structural equation system. There is a specific focus on younger people given the greater impact interventions can have over time. University students are a special group of young people with their own characteristics. Most of them have some financial experience and many will soon be entering the workforce. They are also anticipating managing their student debt, budgeting their household expenditures or saving to purchase a home, and planning for their retirement. University students are therefore an interesting group of young people who can help us better understand what influences the financial behaviors of the younger generation.

Financial behavior
According to Xiao (2008), financial behavior is defined as " . . . any human behavior that is relevant to money management." The most common financial behaviors noted in the literature are those related to the use of money, credit and savings (Hilgert et al., 2003;Xiao et al., 2006). This is why financial literacy is seen as a way to encourage healthy financial behaviors that lead to better financial security in retirement (Kadoya et al., 2018;Lusardi & Mitchell, 2011a, 2011bLusardi & Mitchelli, 2007). While financial literacy is assumed to encourage desirable financial behaviors and improve financial wellbeing (Xiao & Porto, 2017), many individuals appear to lack the financial literacy required to make appropriate and cost-effective financial decisions (Mandell & Hanson, 2009).

Financial confidence
According to Palameta et al. (2016), financial confidence is the self-assurance required to make sound financial decisions. While financial knowledge carries obvious benefits in terms of positive financial behavior, researchers have suggested that individuals who demonstrate higher financial confidence may be better able to implement healthy financial choices (Atlas et al., 2019;Hilgert et al., 2003;O. Stolper, 2018). An interesting perspective was adopted by Atlas et al. (2019), who argued that the impact of financial knowledge on financial decision-making and subsequent financial behavior appears to be short-lived and is particularly dependent upon financial confidence. Therefore, we hypothesize that financial confidence positively influences financial behavior (H1).

Financial knowledge
According to several authors, adopting adequate financial behavior depends largely on one's financial knowledge (Allgood & Walstad, 2016;Mountain et al., 2020). Individuals who have the ability to use their knowledge to make appropriate decisions for the effective management of their financial resources tend to be more financially literate (Asaad, 2015 andWoodyard (2011).. In this regard, numerous studies have assessed the level of financial knowledge among students, young adults and the general population. Some authors have chosen the widely used Jump$tart questionnaire (Keown, 2011;Lalonde & Schmidt,) while others have developed similar questions that relate to a body of financial knowledge or to some of its components (Allgood & Walstad, 2016;Kim et al., 2019Kim et al., , 2020Potrich & Vieira, 2018;Shen et al., 2016). In all cases, the results are the same: financial knowledge is relatively poor among participants. While this has led several authors to take an interest in financial knowledge as a means of improving financial behavior, conflicting results have been obtained. For example, Xiao et al. (2011), Allgood and Walstad (2016), and Mountain et al. (2020) found a positive relationship between financial knowledge and financial behavior, while Borden et al. (2008), and Tang et al. (2015) found that financial knowledge is not sufficient to change financial behavior. Despite these contradictions, we hypothesize that financial knowledge positively influences financial behavior (H2a) and financial confidence (H2b).

Learning capacity
Beyond knowledge and confidence, it is crucial for individuals to understand that they have primary responsibility for acquiring new information and skills in order to effectively handle their personal finances (Loibl & Hira, 2005). Indeed, rapid and continuous changes in financial conditions and opportunities continue to place a heavy burden on individuals' adaptive capacity and resources (Jarvela, 2006). As learning is at the core of such a process (Jarvela, 2006), we argue that an individual's learning capacity can serve as a valuable resource that leads to appropriate financial behavior. In other words, we suggest that having good financial behavior requires a minimum level of financial knowledge and individuals must often acquire this on their own. It therefore seems appropriate to verify whether the ability to acquire this knowledge affects financial behavior. Thus, we argue that individuals who have greater learning capacity will have better financial behavior (H3a), confidence (H3b) and knowledge (H3c).

Sociodemographic variables
To better understand how financial knowledge, confidence and learning capacity influence financial behavior, we also consider the impact of age, gender, education, and ethnicity on these constructs and variables.
In terms of age, studies show that life experience and age improve financial knowledge (Chen & Volpe, 1998;LaBorde et al., 2013). In line with this, we hypothesize that financial knowledge increases with age (H4a).
In terms of gender, a certain number of studies reveal that women tend to have worse financial behavior than their male counterparts (De Bassa Scheresberg, 2013;Chen & Volpe, 1998) and are more at risk of having low financial literacy scores (Tamborini et al., 2012). In this regard, we hypothesize that men will have better financial behavior (H5a), confidence (H5b), knowledge (H5c) and learning capacity (H5d) than their female counterparts.
In terms of education, many researchers have indicated that high school, college or university financial education can improve financial knowledge (Chen & Volpe, 1998;Chung & Park, 2014;Walstad et al., 2010;Xiao et al., 2011). However, contrary results have been confirmed in a metaanalysis conducted by Fernandes et al. (2014), suggesting a limited impact of financial education on financial behavior. In this study, we focus on financial education at both the high school and university levels, distinguishing between those who have taken a university course in a field related to finance and those who are in the business field. We hypothesize that taking a course in a field related to financial literacy at the high school or university level or being in the business field at the university level will improve financial behavior (H6a), confidence (H6b) and knowledge (H6c). We will verify these three hypotheses separately for each aforementioned education level.
In terms of ethnic background, some studies have demonstrated a link between financial knowledge, race and ethnicity (Lusardi et al., 2010). For example, Lusardi et al. (2010) found that Caucasians are more likely than African Americans and Hispanics to correctly answer financial-literacy questions. Assuming that financial knowledge leads to better financial behavior, we hypothesize that Caucasians will have better financial behavior (H7a) and confidence (H7b) and a higher level of financial knowledge (H7c).
Based on the literature discussed above, we propose the hypothesized model shown in Figure 1.

Sample and data collection
During spring 2014, we carried out a self-report survey among Canadian university students to examine their financial knowledge, confidence, learning capacity and behavior. From a list of Canadian universities obtained on the internet, an email that included information on the study and an invitation to complete an anonymous survey, on the SurveyMonkey online platform, was sent to the administrative secretaries of 81 faculties of Canadian universities.
The universities were randomly selected to cover all Canadian provinces. The recipients of this email then had the option of forwarding the research invitation to their universities 'students. A $ 500 prize was offered to increase the participation rate. A total of 1,221 usable surveys were received from across all provinces and territories in Canada. The questionnaire instrument was approved by the Ethical Research Committee of the University of Moncton.

Survey design
The survey included 206 questions that measured self-reported financial knowledge, financial confidence, learning capacity and financial behavior, as well as 6 sociodemographic questions relating to age, gender, education level and ethnicity.

Financial knowledge
In this study, there are 25 questions that measure various components of financial literacy, such as income, savings, investments, spending, credit and money management. These questions are inspired by the Jump$tart questionnaire (Lusardi & Mitchelli, 2007;Mandell, 2008) and have been adapted to the Canadian context. They are similar to those used in other studies, which also attempt to measure financial knowledge (Allgood & Walstad, 2016;Kim et al., 2019Kim et al., , 2020Mountain et al., 2020;Potrich & Vieira, 2018;Shen et al., 2016). The results for these questions were compiled and each respondent was assigned a score. The overall mean score for all participants in this study was 53%.

Financial confidence
The questions used to measure financial confidence were taken from those presented in Robb and Woodyard (2011). Participants answered four questions on a Likert-type scale ranging from 1 to 5. Sample questions included, "I am good at dealing with day-to-day financial matters such as checking accounts, credit and debit cards, and tracking expenses" and "I am pretty good at managing my finances." The Cronbach's alpha for this scale was 74.7.

Learning capacity
The survey questions that aimed to assess individuals' abilities to acquire relevant financial knowledge were adapted from Hedlund et al. (2003) and Smith (2001). Participants answered three questions on a Likert-type scale ranging from 1 to 5. The sample questions included, "I have the aptitude to process a volume of information that is superior to what an average person could process." and "I have a superior ability to consider solutions and seize opportunities that come my way". The Cronbach's alpha for this scale was 74.9.

Financial behavior
The questions that aimed to assess financial behavior were adapted from Fernandes et al. (2014). Participants answered 6 questions on a Likert-type scale ranging from 1 to 5. Sample questions included, "I follow a careful monthly budget" and "I put money aside on a regular basis for future needs". The Cronbach's alpha for this scale was 79.3.

Sociodemographic variables
Our survey included sociodemographic questions that assessed gender, age, ethnicity and educational background. In this study, 74% of participants are female, the average age of university students in our sample is 25 years old, 85% are Caucasian, 9% are in a business discipline, 71% had taken a course related to financial literacy at the university level and 48% had followed one finance-related course in high school.

Results
The descriptive statistics for the main study variables are reported in Table 1. The adjustment indices of the theoretical model are reported in Table 2, which shows that the adjustment of the initial measurement model was insufficient. In fact, the root mean square error of approximation (RMSEA) was not less than 0.05 and the comparative fit index (CFI) and Tucker-Lewis index (TLI) were not more than 0.95. Therefore, the indicators did not allow us to validate the model. In order to correct these shortcomings, we used the Lagrange indicator to check for any possible modifications that would improve the model fit. By adding correlation coefficients between certain indicators, we noted that the final measurement model was considerably improved. In fact, the RMSEA approached a threshold of 0.05 and the CFI and TLI were both above 0.95. In addition, the weighted root mean square residual (WRMR) decreased considerably. Finally, we added links between the different factors in order to arrive at the structure model. Overall, we observed a well-adjusted model with an RMSEA of 0.044. However, the chi-square indicated that the fit of the data to the model was not optimal, the latter being significant. This was expected, however, given the large sample size. Indeed, the chi-square test is a very sensitive adjustment measure to the sample size (Byrne, 2013), and there is a good chance that an adequate model would be rejected in large samples (Kyriazos, 2018). While the chi-square statistic is no longer used as a basis for acceptance or rejection (Schermelleh-Engel et al., 2003), we did respect the rule that the Chi-square divided by the number of degrees of freedom should be less than 5 (Kyriazos, 2018). Indeed, we obtained values of 4.67 and 3.74 for the final measurement model and the structure model, respectively.    Table 3 presents the final model structure, allowing us to verify our hypotheses. In part A of Table 3, our results indicate that financial confidence is positively linked to financial behavior, which is in agreement with (H1) and consistent with the results of previous research (Asaad, 2015;Bannier & Schwarz, 2018;Robb et al., 2012). However, contrary to H2a, financial knowledge does not influence financial behavior. This result is consistent with Fernandes et al. (2014), who conducted a meta-analysis and found a limited impact of financial education on financial behavior. Other researchers have also found that financial knowledge is not enough to influence financial behavior (Borden et al., 2008;Mandell & Klein, 2009;Tang et al., 2015). Our study results also reveal that learning capacity does not influence financial behavior, contrary to (H3a). In part B of Table 3, our results indicate that financial knowledge (H2b) and learning capacity (H3b) influence financial confidence. In part C of Table 3, our results reveal that learning capacity influences financial knowledge, as expected in (H3c).
Regarding sociodemographic variables, our results reveal that age has a positive influence on financial knowledge, which conforms to hypothesis (H4a). With regard to gender, the results indicate that women report better financial behavior (H5a), poorer confidence (H5b) and poorer learning capacity (H5d) than men. The women in the study also obtained lower scores on financial knowledge (H5c) than the men, thus confirming hypotheses H5b, H5c and H5d. These results are consistent with several studies, which also concluded that women have poorer financial knowledge and less financial confidence than their male counterparts (De Bassa Scheresberg, 2013;Bucher-Koenen et al., 2017;Chen & Volpe, 1998, 2002Lusardi & Tufano, 2015).
When we look at the effect of having taken a course in financial management or a related field in high school or university, the results indicate that university courses have a more direct effect on financial behavior than high school courses. Indeed, the results reveal that having taken a course in high school has no influence on financial behavior (H6a), nor does it affect financial confidence (H6b). However, it does have an impact on financial knowledge (H6c). As for university courses, the results show that studying business (BUS) improves financial confidence and financial knowledge (H7b and H7c) and that having followed a university course related to financial field improve financial behavior (H8a). However, having followed a university course related to financial field does not improve financial knowledge or financial confidence (rejection of H8b and H8c). Therefore, the results concerning the influence of education on financial behavior are not unanimous. By comparing our results with previous literature, which also demonstrates contradictory results concerning the influence of education on financial behavior (Hilgert et al., 2003;Mandell & Klein, 2009;Xiao et al., 2011), we argue that course content and student maturity could not only have an important influence on the results but could also possibly explain contradicting results.
Finally, even if being Caucasian improves financial knowledge (H9c), our results do not reveal a direct impact on financial behavior (H9a) or financial confidence (H9b). Other studies have demonstrated that Caucasian students have more financial knowledge (Lusardi et al., 2010) and these results are expected since many non-Caucasian university students are usually international students who are not necessarily familiar with certain Canadian financial concepts. However, the current study shows that their lower financial knowledge does not affect their financial behavior, nor does it affect their financial confidence. Finally, the results shown in Table 3 reveal that the structural equation model can explain 79.3% of the variance in financial knowledge and 34.8% of the variance in financial confidence.

Discussion and implications for researchers and practitioners
Although studies have already established that financial confidence influences financial behavior (Asaad, 2015;Atlas et al., 2019;Bannier & Schwarz, 2018;Robb et al., 2012;O. A. Stolper & Walter, 2017), this study adds to the literature by showing that financial confidence is the construct with the highest explanatory potential for financial behavior, followed by gender and, to a lesser degree, university courses in financial management or a related field. Given the important contribution of financial confidence to an explanation of financial behavior, it is justifiable to question what influences financial confidence.
Our results reveal that, first and foremost, an individual's learning capacity influences their financial confidence, followed by financial knowledge, gender, and being in a business program at university. These results confirm the results of previous research which had shown that the relationship between financial confidence and financial behavior was influenced by other variables such as gender (Bucher-Koenen et al., 2017) and financial knowledge (Atlas et al., 2019). Additionally, our model helps highlight the importance of learning capacity in explaining financial confidence which is an original contribution of this study. To our knowledge, no prior study has evaluated this relationship which, according to our results, should be considered in a conceptual model of financial literacy. Given that our model only explains 34.8% of the variance in financial confidence suggest that future research should focus on providing a better explanation of this factor.
It is important to remember that several studies have already looked at the direct link between financial knowledge and financial behaviour (Allgood & Walstad, 2016;Borden et al., 2008;Mountain et al., 2020) and the results of those studies are sometimes contradictory in the sense that some find that financial knowledge contributes to positive financial behaviour while others do not. The results of this study provide a plausible explanation for this contradiction due to the fact that we obtain an indirect relationship between financial knowledge and financial behaviour rather then a direct relationship (validation of H2b and non-validation of H2a).
Next, our results also suggest that high school financial education should be improved in order to increase its influence on financial behavior. These findings are consistent with those of Fernandes et al.'s (2014) meta-analysis, where they found a limited impact of financial education on financial behavior and financial capability and suggest that financial education is necessary and should be improved to increase its effectiveness.
In light of all these results, certain implications for educators, government organizations, advisors in financial institutions and academicians should be considered. First, strategies that aim to improve financial literacy should consider the crucial role of financial confidence in explaining financial behavior and the indirect influence of learning capacity on financial confidence. To a lesser degree, age, gender, financial knowledge, education and ethnicity are also variables that seem to influence financial behavior directly or indirectly. This information is particularly relevant for financial advisors. Indeed, knowing the extent to which these variables influence financial behavior, financial confidence and financial knowledge could help advisors be more efficient in their role, which is to mitigate existing financial problems, prevent new ones from occurring and implement constructive strategies to achieve specific financials goals (Haman & Laker, 2018).

Conclusion
This study used a structural equation model and a large national dataset to examine the impact of financial knowledge, confidence, learning capacity, and socio-demographics on the financial behaviour of university students. The results show that the hypothesized model explains 79.3% of the variance in financial behaviour among the participants in our sample.
This study contributes to the literature by evaluating the impact of a unique set of constructs and variables of financial behaviour and the interrelationships between these variables so as to allow a better understanding of how they influence financial behaviour. The results of this study are relevant for the various stakeholders who are interested in improving financial literacy, they also complement other studies in the field of financial literacy.
The main results of this study highlight the importance of financial confidence in influencing people to adopt better financial behaviours. Findings also point to the influence of learning capacity and financial knowledge on financial confidence.
The current study has a few limitations that should be addressed in future research. First, it used cross-sectional data that can only document associations between financial knowledge, confidence, learning capacity and financial behaviour. Future research could use longitudinal data to verify the direct and indirect relationships between these variables. Another limitation is in the limited set of constructs and variables we used to explain financial behaviour, as this set does not give us a complete picture of the actual phenomenon. Therefore, we suggest researchers examine additional and complementary psychological determinants of financial behaviour to eventually be able to better conceptualize the phenomena associated with financial behaviour.