Defining the Predictors of Financial Literacy for High-School Students

Abstract

This work documents the significant predictors of financial literacy of high-school students across Greece. The results derived, from the examined aspects of financial knowledge, financial behaviour, and financial attitude, the determinants of financial literacy that high-school students demonstrated. A novel approach whereby a canonical analysis is applied to financial literacy data. The findings reveal that better-performing students, students with parents with a university degree, students’ perception of parents’ income, and students who keep records of their income and expenses are significant predictors of financial literacy. The results of this analysis should be considered when building financial education initiatives to embrace the young generation.

1. Introduction

Financial literacy is defined as the combination of awareness, knowledge, skills, attitudes, and behaviours that is conducive to sound financial decisions and ultimately to personal and household financial well-being (Atkinson & Messy, 2012; G20, 2012). With the PISA 2018 measurement, the OECD (2020) refined the definition and gave an enriched version that accommodates students’ capacity to use financial information and their capability to successfully cater to financial challenges in the future. The definition stated that financial literacy was the knowledge and understanding of financial concepts and risks and the skills, motivation, and confidence to apply such knowledge and understanding to make effective decisions across a range of financial contexts to improve the financial well-being of individuals and society and to enable participation in economic life.

Measurement of financial literacy traces back to the works of (Hahn, 2014; Klapper & Panos, 2011; Klapper et al., 2013, 2015; Lusardi & Mitchell, 2006; Mandell, 2008; OECD INFE, 2016a, 2016b; OECD, 2014, 2017, 2020, 2024a, 2024b). These studies indicate the importance of financial literacy for financial well-being and argue that financial literacy is low among the young, making them ideal targets for financial education programmes. Putting the concepts above into the context of national level, financial literacy should be examined, representing a mapping of peoples’ financial knowledge, financial behaviours, and financial attitude. Focus is given to 15-year-old high-school students, as globally, this is the point at which most children are still enrolled in formal education. Hence, according to OECD (2024b), many 15-year-olds face financial decisions and are already consumers of financial services. Therefore, they are likely to face growing complexity and risks in the financial marketplace as they move into adulthood.

Atkinson (2018) states that for a national financial literacy strategy to be fulfilled, it is essential to initially identify the national needs and gaps in relation to financial literacy. To this end, many countries have optionally participated in the OECD’s financial literacy section of Programme for International Student Assessment (PISA) that addresses 15-year-old students every three or four years. The latest assessment was performed in 2022, while the previous ones took place in 2018, 2015 and in 2012 (OECD, 2014, 2017, 2020, 2024b). The surveys evaluated the knowledge and skills of 15-year-old students with respect to money matters and personal financial knowledge. The topics covered entailed the understanding of bank accounts and debit cards, understanding interest rates on a loan, and choosing between a variety of mobile phone plans, inter alia. In 2022 and 2018, 20 countries participated, while in 2015 and 2012, 15 and 13 participated, respectively (OECD, 2014, 2017, 2020, 2024b). However, Greece has never participated in the optional Financial Literacy Module of PISA. Only recently, Greece participated at the OECD survey on adults and young people, as part of Greece’s national financial literacy strategy (OECD, 2024a). As only recently, Greece’s national strategy of financial literacy was kicked-off, only a few studies can be found. The most recent one is on the financial capability of 15-year-olds in Greece (Tzora et al., 2023). While in a prior study in 2019, Tsakiridou and Seitanidis (2019) surveyed 15-year-old students in Thessaloniki, Northern Greece and Tzora (2025), and Philippas and Avdoulas (2019) surveyed university students in Athens, Central Greece.

OECD (OECD INFE, 2016a; OECD, 2024a) developed a theoretical model for the overall levels of financial literacy (indicated by combining scores on knowledge, attitudes, and behaviour) that analyses the characteristics of participants in the survey. They argue that financial literacy levels are lower than may be expected for a variety of reasons that, in each case, need to be elaborated.

The study of financial literacy among high-school students in Greece is of great importance for several reasons. Firstly, secondary education, up to age 15, is compulsory in Greece, and it is free in public schools. According to the Hellenic Statistical Authority (2018), there were 1814 junior high schools in Greece in the year 2015–2016. All schools follow a national curriculum instructed by the Hellenic Ministry of National Education and Religious Affairs. In the school curriculum, a personal finance course that meet their everyday needs is missing. Instead, students are learning general home economics early in junior high school. During the last year of high school, students who wish to take the panhellenic exams aiming for an economics-related discipline are taught a much more specialized economics course, involving an introduction to microeconomics and macroeconomics. This gap in financial literacy might turn out to be vital for the students’ financial well-being. Greek students may need to make personal financial decisions both when they finish their studies with just the compulsory education and when they continue their studies in high school and University. Students who pursue a higher education after high school make various financial decisions every day and may face challenging choices on how to finance their studies or support themselves through their studies. Although undergraduate studies in public universities in Greece are free, many students choose to study away from home and need to handle their financial affairs by themselves for the first time.

Secondly, in the of post-pandemic period and times of long-term economic crisis, the challenges to financial resilience and financial well-being among the youngest generation are likely to be exacerbated. In this context, financial education can play a pivotal role in properly managing personal financial affairs and protecting people from potential financial shocks. Moreover, the shrinking of government programmes to people in need and the increased advancement and development of available financial tools and services have all added to a more prominent awareness that ensuring people, especially young people, are financially literate is important (OECD, 2012).

Also, many people have encountered e-commerce through various platforms, and this experience made them vulnerable to these new digital activities, such as financial fraud. The Stanford Center on Longevity (2018) reported that younger people lose money to fraud more often than older people. The long-term economic crisis will further increase the existing inequalities between the generations. Panos and Wilson (2020) reported that in an era of increased (digital) financial inclusion and threats arising from instances of (online) financial fraud, financial education and financial advising enhance financial and overall well-being.

Given the above, it is argued that, now more than ever, citizens need to be aware of financial matters because of the long-term economic crisis and its consequences for their financial resilience and financial well-being. Last, OECD (2024b) reported that many 15-year-olds are already consumers of financial services and are likely to face risks in the financial marketplace as they move into adulthood.

Through the survey, information was collected on demographics and socioeconomic characteristics of students’ financial knowledge, behaviours and habits, and attitudes. These characteristics were measured in accordance with the proposed analysis by OECD (OECD INFE, 2016a; OECD, 2024a). The question design stemmed from a thorough inquiry into the existing literature (Hahn, 2014; Klapper & Panos, 2011; Klapper et al., 2013, 2015; Lusardi & Mitchell, 2006; Mandell, 2008; OECD, 2014, 2017, 2020, 2024a, 2024b; OECD INFE, 2016a, 2016b), which were adjusted to Greek standards and to meet the reality.

Among the others, the determinants include these students’ characteristics (socio economical, family, etc.) that lead to acceptable levels of financial knowledge, financial behaviours, and financial attitude.

From a theoretical perspective, this work contributes a novel analysis of canonical analysis of students’ characteristics for financial components. To the best of the author’s knowledge, this is the first time that a study has applied canonical analysis to financial literacy data. Specifically, the cutoff point for an acceptable level of financial knowledge was a minimum of 70 percent proficiency (as per the proposed analysis of OECD, 2024a; OECD INFE, 2016a). Also, the cutoff point for acceptable levels of financial behaviour and financial attitude was the minimum of 70 percent proficiency. Hence, a factor analysis was performed to ensure that the questions measured the relevant category of financial literacy (e.g., financial knowledge, financial behaviour, and financial attitude). These analyses made possible the examination of differences concerning financial literacy.

The remainder of the paper is organised as follows: Section 2 outlines the relevant literature and measurements of financial literacy for the youth that is conducive to validate the objectives and rationale the means of the research. Section 3 presents the survey instrument and dataset that enable the empirical strategy. Also, it presents the sample of the high-school students of the survey. Section 4 presents the descriptive and regional analysis of the students in Greece. Section 5 presents the canonical analysis of the factors that defining the predictors of financial literacy for high-school students in Greece. Section 6 concludes with a discussion that assesses the potential policy impact of the findings and outlines avenues for future research.

2. Background and Literature Review

2.1. The OECD Measurement of Financial Literacy Among 15-Year-Old Students

In the PISA measurements of 2012, 2018, and 2022, on average, only one in ten students reached the highest level of financial literacy across OECD economies. However, in the PISA 2015 survey, this number was slightly better. In PISA 2015, less than 5 percent of the participant students in each country reported that they did not know what a bank account was. In PISA 2018, roughly one in three students had the ability to comprehend and evaluate a bank statement. In another interesting finding, the analysis of the data from PISA 2015 disclosed that immigrant students scored 26 points lower on average in an assessment of financial literacy than native-born students of similar socioeconomic status, while that number was further increased to 30 points in PISA 2018. Thus, in the PISA 2022 assessment, focus was given to parents’ involvement. Results revealed that on average, across OECD countries and economies, 83% of students reported talking to their parents at least once a month about money for things they want to buy and 76% about their own spending decisions (OECD, 2014, 2017, 2020, 2024b).

The findings of all PISAs regarding gender differences disclosed small to zero differences in financial literacy scores. However, there were gender differences in terms of attitude and behaviour. Specifically, gender differences for high-school students in PISA 2012 showed that there were no differences in average financial literacy scores in all participating countries except Italy, where male students scored better than female students. In PISA 2015, there were no gender differences in scores in nine out of fifteen participating countries. Only in five countries (Australia, Lithuania, Poland, Slovak Republic, and Spain) did female students score higher on financial literacy. Again, in Italy, male students scored higher than female students on the PISA 2015. In PISA 2018 and PISA 2022, there was no significant difference in performance between the male and female students on average across all countries that participated in the assessment. However, males scored higher on financial literacy than girls in the PISA 2018 and in PISA 2022 on average across OECD countries.

2.2. Other Measurements of Financial Literacy Among Students

Apart from the OECD PISA assessments on financial literacy, various studies have determined the level of financial literacy in high-school students. Students’ comprehension of financial concepts and their level of financial literacy is an important factor in determining how they control their money and overcome financial obstacles. A very well-known measurement tool that various researchers (McKenzie, 2009; Sohn et al., 2012; Thaden & Rookey, 2004) and the Federal Reserve Bank in America have used to measure the level of financial literacy (Hilgert et al., 2003) is the Jump$tart survey (Mandell, 2009). They have conducted this survey for 10 years among high school seniors in the United States. The 2008 survey results (Mandell, 2008) disclosed that high school seniors who planned to attend a four-year college did much better on the measurement than other seniors. It found that only 16.8 percent of high school seniors felt that stocks were likely to have higher average returns than savings bonds, savings accounts, and checking accounts over an 18-year period and only 27.3 percent of high school seniors realized that interest on a savings account was taxable if one’s income was high enough. Moreover, about 40 percent of high school seniors realized that their health insurance could stop if their parents became unemployed.

A survey of 1416 high-school students in Germany showed that students who performed better in elementary school and students with mathematical skills had higher levels of financial literacy (Erner et al., 2016). Cameron et al. (2014) conducted a study on 352 high-school students from five schools in New Zealand and found that students who were financially poorer, students with less English ability, and students with less mathematical ability had lower scores on financial literacy. The results from India (Jayaraman & Jambunathan, 2018) showed that students, despite having high levels of numeracy, were unable to transfer that knowledge to financial computations. The survey of Ghazali et al. (2017) on 458 high-school students in Malaysia showed that students with business knowledge had higher levels of financial literacy than students without that knowledge. They claimed that the knowledge had significant effects on their understanding of certain financial literacy issues. Khoirunnisaa and Johan (2020) conducted a study in Indonesia and found that students with science majors had higher levels of financial literacy than students with social science majors. The results also showed that students with high grades had a better understanding of financial concepts compared to students with low grades.

Arceo-Gómez and Villagómez (2017) surveyed 889 high-school students in 16 Mexican high schools and found that less than one in five students was in a position to understand basic financial concepts. They also found that school type (private or public) and household income were not essential components in enhancing financial literacy. Arceo-Gómez and Villagómez (2017) found similar results to PISA, where there were no generalized gender differences in Mexico. However, they also found that female students scored a bit higher mostly due to better financial behaviour. Erner et al. (2016) conducted a study in Germany and found that female high-school students scored lower than male students. By contrast, Ghazali et al. (2017), Jayaraman and Jambunathan (2018), and Liu et al. (2019) reported in their work in Malaysia, India, and in Taiwan, respectively, that females outperformed males in high school.

In the next Section, the survey instrument and data used in the study are presented, to complement the existing evidence, by assessing the financial literacy of 15-year-olds in Greece and defining the predictors of financial literacy.

3. The Data

3.1. The Survey

The questions used in the survey instrument were derived from an extensive literature review to measure the levels of financial literacy in Greece. To this extent, the survey included seventeen questions on financial knowledge, seven questions on financial behaviours, and three questions on financial attitudes. Although the questions had been used in various surveys in the past (Hahn, 2014; Klapper et al., 2015; Mandell, 2008; OECD INFE, 2016a; OECD, 2024a), some were adjusted to meet Greek standards and meet the specifics and contextual realism of Greece. The below Figure (Figure 1) presents in detail the components of financial literacy used as the survey instrument.

The following Table (

Table 1. Component of financial literacy.

Financial Knowledge (FK)

Q.FK.1-Diversification

Q.FK.2-Interest Rate

Q.FK.3-Inflation

Q.FK.4-Compound Interest

Q.FK.5-Product Offer

Q.FK.6-Discount

Q.FK.7-Tax

Q.FK.8-Pension

Q.FK.9-Insurance

Q.FK.10-Salary

Q.FK.11-ATM

Q.FK.12-Reading Bill (a)

Q.FK.13-Reading Bill (b)

Q.FK.14-Share chart

Q.FK.15-Selling share

Q.FK.16-Period of saving

Q.FK.17-Safe keeping

Financial Behaviour (FB)

Q.FB.1-Saving short term

Q.FB.2-Buying

Q.FB.3-Comparing prices

Q.FB.4-Dealing with money

Q.FB.5-Managing finance affairs

Q.FB.6-Covering expenses

Q.FB.7-Saving long term

Financial Attitude (FA)

Q.FA.1-Liking economics

Q.FA.2-Financial Knowledge

Q.FA.3-Making money

) presents the description of the questions included in each component of financial literacy (i.e., financial knowledge, financial behaviour, and financial attitude).

In the current study, all the questions of the measurement tool are used. The increased number of questions in financial knowledge component of financial literacy, as per OECD recommendations (OECD, 2024a) reflects the need for testing financial literacy through more knowledge-based questions. Therefore, this study assesses financial knowledge questions more than financial behaviour and financial attitude as per OECD recommendations.

It is worth mentioning that the questionnaire was approved by the Hellenic Ministry of Education, Research, and Religious Affairs. In Greece, to survey high-school students requires an approval by that Ministry. A dedicated committee is assigned to evaluate the questionnaire to exclude any questions that may discomfort students in need (e.g., sensitive family financial matters).

The dataset was collected via a customized online survey questionnaire. The access to the junior high-school population was granted by the Ministry of Education, Research, and Religious Affairs under approval number 41,396/Δ2 dated 9 March 2016. The sample was designed to be nationally representative (via proportional stratified random sampling), and the approval from the ministry was for contacting 260 high schools. In total, 94 out of the 260 schools responded to the invitation, which resulted in a response rate of 37 percent. As a result, 3529 15-year-old students were invited to participate in the online survey. Among the responses received, 3028 were returned completed and usable for our analysis.

Since the study did not use a single measurement scale for financial literacy but rather a combination, the selected items of the study were subject to an exploratory factor analysis (Field, 2013). The factor analysis was performed for 27 items that measure financial literacy. Based on Kaiser–Meyer–Olkin (KMO) and Bartlett’s tests, the data were adequate to perform the principal component analysis (Field, 2013; Hair et al., 2019). The factor analysis results are available in the results in Section 5.1.

The approach adopted is the recommended one of OECD (OECD INFE, 2016a; OECD, 2024a), where it measures the elements (financial knowledge, financial behaviours, and financial attitudes) that comprise financial literacy. Hence, a student is financially literate if they score 70 percent or over. Thus, 42.5 percent of the sample scored at or above the threshold, with the remaining 57.50 percent scoring below the threshold.

3.2. The Sample

The demographic characteristics of the sample are presented in

Table 2. Characteristics of the sample.

Variables		Frequency	Variables		Frequency	Variables		Frequency
Gender	Male	49.10%	Repeat a school year	No	96.60%	Mother’s Education	No education	1.30%
	Female	50.90%		Yes	3.40%		Primary School	4.50%
Nationality	Greek	92.70%	Type of High School	Private	5.10%		Lower High School	9.40%
	Other	7.30%		Public	94.90%		Upper High School	28.70%
Region of Greece	Eastern Macedonia and Thrace	9.05%	Classification of High School	Classical	93.90%		Post-secondary education	13.80%
	Attica	24.14%		Art	0.20%		BA/BSc	30.40%
	North Aegean	2.74%		Musical	1.60%		MA/MSc	7.60%
	Western Macedonia	3.70%		Experimental	4.30%		PhD	4.30%
	Western Greece	8.72%	Father’s Education	No education	1.30%	Graduation Grade	Average (10–12)	2.30%
	Epirus	3.24%		Primary School	6.20%		Almost Good (12.1–14)	7.00%
	Thessaly	8.86%		Lower High School	12.90%		Good (14.1–16)	17.60%
	Ionian Islands	3.40%		Upper High School	27.80%		Very Good (16.1–18)	28.00%
	Central Macedonia	11.53%		Post-secondary education	15.30%		Excellent (18.1–20)	45.10%
	Crete	2.70%		BA/BSc	25.40%
	South Aegean	5.55%		MA/MSc	6.90%
	Peloponnese	10.96%		PhD	4.20%
	Central Greece	5.41%
Knowledge of Income Living	No	53.80%	Pocket Money	I am not receiving	16.80%	Participate at an Economic Competition	No	39.20%
	Yes	46.20%		Now, I am not receiving	1.40%		Yes	60.80%
Living	Two Parents	85.10%		€0–€5	21.10%	Wish to have a class of personal finance in High School	As a new subject in school	38.20%
	Mother	11.40%		€6–€10	28.90%		As a part of another subject in school	32.3%
	Father	1.60%		€11–€20	19.60%		As an optional subject but NOT in school	21.00%
	Other	1.9%		€21–€30	6.60%		I am not interested	8.50%
Swift/Change in Family Income	Reduced	80.10%		Over €30	5.60%	Keep records of income/expenses	No	76.10%
	Stable	18.00%	Swift/Pocket Money	Reduced	33.00%		Yes	23.90%
	Increased	1.90%		Increased	5.20%
				Stable	61.80%

. The sample consists of 50.9 percent female and 49.1 percent male students. Most of the students are of Greek nationality (92.7 percent), with an additional 7.3 percent having a different nationality. With respect to the 13 administrative regions of Greece, 24.1 percent reported living in the region of Attica, while 11.5 percent reported living in the region of Central Macedonia, and 11 percent reported living in the region of Peloponnese. Students also reported the exact prefecture they lived in.

Most of the students (94.9 percent) attended a public school, while the remaining students (5.1 percent) attended a private school. Furthermore, students reported if their school was classified as an “experimental”, “music”, “art”, or a “classical” school. Most students (93.9 percent) in the survey attended a “classical school” while a few attended an “experimental school” (4.3 percent) or a “music school” (1.6 percent). In total, 45.2 percent of the students reported exemplary attainment of the learning outcomes of their courses, i.e., obtaining a grade between 18.1 and 20 out of 20 in the previous school year. A total of 28 percent of the sample received a grade denoting conclusive attainment of the learning outcomes, i.e., 16.1–18 out of 20. The remaining 26.8 percent exhibited variable attainment of the learning outcomes of the previous year in their school.

Students also reported their parents’ educational level. The majority (27.8 percent) reported that their father graduated from an upper high school (lyceum in Greece), while 25.4 percent reported that their father held a university degree (undergraduate level). In total, 6.9 percent and 4.1 percent reported their fathers as having a master’s and a Doctoral degree (PhD), respectively. With respect to their mother’s education, 30.4 percent reported that their mother held a university degree (undergraduate level). This percentage is significantly higher than the equivalent percentage of the father’s educational level. Thus, more mothers than fathers were graduates with a master’s and a PhD degree, i.e., 7.6 percent versus 4.3 percent and 6.9 percent versus 4.1 percent, respectively. The second highest percentage of mother’s educational level is 28.6 percent graduating from upper high school.

Furthermore, students were asked to report if they were aware or not of their family income and to give an estimation of the family’s changes in their income monthly. In the sample, most of the students (53.8 percent) stated that they were not aware of their family’s monthly income. However, 2058 students (80.1 percent) claimed that their family’s monthly income had decreased due to the financial crisis, regardless of whether they were or were not aware of the amount. Also, only 18 percent of the students claimed that their family’s monthly income remained stable. Regarding their pocket money (allowance), most students (61.8 percent) reported that their amount remained the same, while 33 percent stated that their amount decreased due to the financial crisis. An additional 1.4 percent (42 students) reported that they were now not receiving pocket money, while they did before the financial crisis. In addition, a variable that stood out in our analysis was whether the students kept or did not keep records of their income and expenses. In the sample, most of the students (76.1 percent) stated that they were keeping records of income and expenses.

Moreover, the students in the survey had to report if they wished to attend a personal finance class during their school hours or not. In total, 38.2 percent of the respondents reported that they wanted to have a personal finance course as a new course, while 32.3 percent of students replied that the personal finance course should be part of an existing course. Finally, students were asked whether they wanted to participate in an economic contest (e.g., Economic Olympiad) with their high school, and 60.8 percent responded positively.

4. Descriptive and Regional Analysis of Student Financial Literacy in Greece

Regional Differences in Financial Literacy

Greece has 13 administrative regions, all of which are represented in this study. Figure 2 presents an overview of the levels of financial literacy in each region. Figure 3 presents the ranking of administrative regions with respect to the levels of financial literacy. The inspection of the Figures indicates that the financial literacy levels are uneven across the country, with a wide range of outcomes. For example, Athens has the highest level at 49.2 percent, Central Greece at 48.5 percent, North Aegean at 43.4 percent, while the lowest levels of financial literacy are in the Ionian Islands (26.3 percent) and Epirus (29.4 percent). The results show that the urban areas and their surroundings have higher acceptable levels of financial literacy. The region of Athens has almost one-third of the population (almost 40 percent) and is the capital of Greece (Hellenic Statistical Authority, 2014)1. In contrast, remote areas and their surroundings have lower acceptable levels of financial literacy. For example, the population at the Ionian Islands region is low compared to Athens and is less than 2 percent of the population in Greece (Hellenic Statistical Authority, 2014).

The regional analysis is enriched by performing correlation analyses. The first analysis revealed that financial literacy levels and regions are statistically significant as determined by one-way ANOVA (F(12, 2909) = 4.825, p < 0.05,

Table 3. One-way ANOVA.

ANOVA
%FinLit
	Sum of Squares	df	Mean Square	F	Sig.
Between Groups	13,743.262	12	1145.272	4.825	0.000
Within Groups	690,497.269	2909	237.366
Total	704,240.531	2921

).

Hence, a Pearson correlation coefficient was computed to assess the linear relationship between financial literacy levels and of the gross domestic product (GDP) per capita per region of Greece (r = 0.091, p < 0.05,

Table 4. Pearson correlation.

Correlations
		%FinLit	RegPerC
%FinLit	Pearson Correlation	1	0.091 **
	Sig. (2-tailed)		0.000
	N	2922	2922
RegPerC	Pearson Correlation	0.091 **	1
	Sig. (2-tailed)	0.000
	N	2922	3028

**. Correlation is significant at the 0.01 level (2-tailed).

). There was a positive correlation between the two variables. This suggests that an increase in the financial literacy levels will increase in the gross domestic product (GDP) per capita per region.

At this point, no more regional analyses will be performed, as this work aims to examine the sample as a whole, at a national level, since it is representative to the young population in Greece.

5. Multivariate Analysis of the Factors Influencing Financial Literacy

In this Section, the theory and the results of the multivariate analysis of the determinants of financial literacy among 15-year-olds in Greece are presented. The aim is to extract information from a large number of variables in order to interpret the relationships between the variables (Carlis, 2005). In line with common practice, the present study introduced the components of financial literacy (financial knowledge, financial behaviour, and financial attitude). Considering the aim, the present research performed descriptive analyses and a distinctive type of multivariate analysis, namely canonical analysis. The results of the analysis are presented below. Regarding the latter, exploratory and confirmatory factor analyses were conducted to effectively summarize the information of the financial literacy elements/items into a set of composite variables or factors (Hair et al., 2019).

5.1. Factor Analysis

Since the present study did not use a single measurement scale for financial literacy but rather a combination, the selected items of the study were subject to an exploratory factor analysis. To be exact, the aim of performing the factor analysis was to examine whether the questions asked were related to the construct that was intended to be measured (Field, 2013).

The most suitable approach for this study was the exploratory factor analysis and, in particular, the principal component analysis (PCA) (Field, 2013). The factor analysis was performed for 27 items that measure financial literacy. Since financial knowledge, financial behaviour, and financial attitude should be correlated (OECD, 2024a; OECD INFE, 2011, 2016a), then oblique rotation should be performed with the factor analysis (Hair et al., 2019; Field, 2013; Yong & Pearce, 2013).

The KMO (Kaiser–Meyer–Olkin) (

Table 5. ΚΜO and Bartlett’s tests.

Kaiser–Meyer–Olkin and Bartlett’s Tests
Kaiser–Meyer–Olkin Measure of Sampling Adequacy	Bartlett’s Test of Sphericity
	Approx. Chi-Square	Df	Sig.
0.852	6956.389	351	0.000

) measurement of sampling adequacy showed an acceptable value of KMO, 0.852, that confirmed the adequacy of the measuring tool for the analysis by being above the acceptable level of 0.5 (Field, 2013). As per the acceptance guide of Hutcheson and Sofroniou (1999), the value of KMO falls within the “meritorious” range of values (0.8–0.9). Also, the Bartlett’s test of sphericity was statistically significant (p < 0.05). Based on these tests, the data were adequate to perform the principal component analysis (Field, 2013; Hair et al., 2019).

Table 6. Summary of the results of the factor analysis (N = 3028).

	Rotated Factor Loading
Item/Question	Financial Knowledge	Financial Behaviour	Financial Attitude
Q.FK.12-Reading Bill (a)	0.615	−0.047	0.006
Q.FK.13-Reading Bill (b)	0.397	−0.062	0.110
Q.FK.5-Product Offer	0.357	−0.030	0.095
Q.FK.6-Discount	0.470	−0.056	0.098
Q.FK.7-Tax	0.531	−0.035	−0.026
Q.FK.8-Pension	0.532	0.051	−0.106
Q.FK.9-Insurance	0.403	0.037	−0.068
Q.FK.10-Salary	0.445	0.013	−0.094
Q.FK.11-ATM	0.589	−0.055	−0.002
Q.FK.16-Period of saving	0.547	−0.041	0.049
Q.FK.17-Safe keeping	0.385	0.137	0.018
Q.FK.14-Share chart	0.504	0.023	−0.045
Q.FK.15-Selling share	0.396	0.068	0.005
Q.FK.1-Diversification	0.264	0.026	0.262
Q.FK.2-Interest Rate	0.287	0.034	−0.013
Q.FK.3-Inflation	0.444	0.001	0.099
Q.FK.4-Compound Interest	0.251	0.017	0. 005
Q.FB.1-Saving short term	−0.044	0.632	0.045
Q.FB.2-Buying	0.178	0.460	−0.164
Q.FB.3-Comparing prices	0.145	0.452	−0.129
Q.FB.4-Dealing with money	0.020	0.530	0.023
Q.FB.5-Managing finance affairs	−0.108	0.374	0.173
Q.FB.6-Covering expenses	0.006	0.453	0.098
Q.FB.7-Saving long term	−0.125	0.602	0.032
Q.FA.1-Liking economics	−0.049	0.041	0.714
Q.FA.2-Financial Knowledge	0.190	0.047	0.606
Q.FA.3-Making money	−0.070	−0.038	0.699
Eigenvalues	4.022	1.731	1.317
Percent of variance explained	14.895	6.413	4.876
Cronbach’s Alpha	0.746	0.530	0.561

Rotation: Promax with Kaiser normalization. Note: Factor loadings greater than 0.30 are in bold.

shows the factor loadings after rotation. Also, it shows the items (questions) that clustered on each factor. The first factor represents the aspects of financial knowledge, the second factor represents the aspects of financial behaviour, and the third factor represents the aspects of financial attitude. The items QFK 1, QFK2, and QFK3 could be excluded, but since they have been extensively used in the financial literacy literature, it has been decided to retain them. Also, eigenvalues are obtained (with over 1 option, Kaiser’s recommendation, (Field, 2013)) for each financial literacy factor in the data.

Based on the results of the factor analysis, the variables used at the study are measuring what they were intended to, since each item (variable) “falls” into the corresponding factor (unit) as defined by the relevant literature review (OECD, 2024a; OECD INFE, 2011, 2016a).

5.2. Canonical Correlation Analysis

The multivariate technique of canonical correlation analysis (CCA) is then performed. As per Hair et al. (2019), the canonical correlation represents the only technique available for examining the relationship between independent variables with multiple dependent variables. By using canonical correlation analysis, it can create a composite measure that consists of all dependent variables. Also, the canonical correlation analysis minimizes the risk of committing a Type I error (Hair et al., 2019; Sherry & Henson, 2005). Therefore, it gives room to examine the ways in which financial literacy and students’ characteristics are related, the strengths of the relationships, and the nature of the relationships (Hair et al., 2019).

The general form of canonical analysis is as follows:

Y1 + Y2 + Y3 +… + Yn = X1 + X2 + X3 + … + Xn
(metric, nonmetric) (metric, nonmetric)

Furthermore, the canonical correlation analysis accommodates metric variables and allows the use of transformed nonmetric data, i.e., in the form of dummy variables (Hair et al., 2019). To this extent, the eight students’ characteristics (region, school classification, graduation grade, father’s education, mother’s education, type of family, change in family’s income, and change in pocket money) are transformed from ordinal (categorical) to binary variables prior to the analysis.

For the application of the canonical correlation, 18 variables are used as the input data. The students’ characteristics questions in the survey are designated as the set of independent variables. The measures of financial knowledge, financial behaviours, and financial attitudes stem from synthetic variables. (a) Q.FK.1 through Q.FK.17, (b) Q.FB.1 through Q.FB.7, and (c) Q.FA.1 through Q.FA.3 are specified as the set of dependent variables. The statistical problem involves identifying any latent relationships between a student’s characteristics and their level of financial knowledge, level of financial behaviour, and level of financial attitude.

By performing this analysis with multiple dependent and independent variables, the author wants to know (a) whether there is a relationship between the predictor and criterion variable sets and, to be more precise, the strength of the relationship between the two sets, (b) what characteristics and financial literacy variables/components are more or less useful in the model, and whether they relate to each other in the expected directions. The measure of the strength of the relationship between the two variates is expressed as a canonical correlation coefficient (R_c). The canonical variates represent the optimal linear combinations of dependent and independent variables, while the canonical correlations represent the relationship between them (Hair et al., 2019).

The analysis yields three (3) functions with squared canonical correlations (R_c²) of 0.232, 0.095, and 0.013 for each successive function, as shown in

Table 7. Measures of overall model fit for canonical correlation analysis.

Measures of Overall Model Fit for Canonical Correlation Analysis
Canonical Function	Canonical Correlation	Canonical R2	F Statistics	Probability
1	0.48	0.232	25.38491	0.000
2	0.31	0.095	11.71512	0.000
3	0.11	0.013	2.77889	0.000

. The full model is statistically significant with a Wilks’s λ = 0.686 criterion, F(45, 8455.52) = 25.385, p < 0.001, as shown in

Table 8. Multivariate test of significance.

Multivariate Test of Significance
Statistic	Value	Approximate F Statistic	Probability
Wilk’s lambda	0.686	25.385	0.000
Pillai’s trace	0.340	24.230	0.000
Hotelling’s trace	0.420	26.544	0.000
Roys	0.232

. Therefore, the analysis shows that at least one correlation is statistically significant and that there probably is a relationship between the two variable sets. Further, by taking 1 − λ, the overall effect for the full model is 1 − 0.686 = 0.314 = R_c².

Because of the size of the canonical correlations in Table 5, the author interprets the first two functions as having canonical correlations of 0.48 and 0.31, respectively. The third function has a canonical correlation lower than 0.30 (0.11), and therefore, the author does not interpret it (Tabachnick & Fidell, 2014). Also, the first two functions explain 23.2 percent and 9.5 percent of the variance within their functions, respectively. Finally, the third function explains less than 9 percent of the variance in the function that is sufficiently weak to not warrant interpretation (Sherry & Henson, 2005).

In summary, the canonical analysis shows that there is certainly a noteworthy relationship among the variable sets of students’ characteristics and elements of financial literacy by means of evidence of statistical significance and effect size. Moreover, this relationship is largely captured through the primary features in the canonical model. Furthermore, this relationship comes largely from the first two functions in the canonical model.

Table 9. Canonical functions.

	Canonical Function 1			Canonical Function 2
Variables	Canonical Weights (w)	Canonical Loadings (rs)	Squared Canonical Loadings (rs2) (%)	Canonical Weights (w)	Canonical Loadings (rs)	Squared Canonical Loadings (rs2) (%)	Commonalities (h2) (%)
Financial Knowledge	0.94	0.99	98%	0.44	0.12	1%	99%
Financial Behaviour	0.09	0.33	11%	−0.93	−0.90	81%	92%
Financial Attitude	0.09	0.41	17%	−0.31	−0.36	13%	30%
R2	42.04%			31.52%
Living in Athens	0.13	0.21	4%	0.08	0.16	3%	7%
Attending public high school	−0.08	−0.19	4%	−0.12	−0.1	1%	5%
Attending classical high school	−0.2	−0.37	14%	−0.13	−0.16	3%	16%
Female	0.1	0.23	5%	0.3	0.37	14%	19%
Greek nationality	0.04	0.19	4%	−0.13	−0.11	1%	5%
Excellent grades in high school	0.59	0.81	66%	0.09	0.12	1%	67%
Repeat a school year	−0.13	−0.23	5%	0.09	0.06	0%	6%
Father’s university-level education	0.18	0.52	27%	−0.07	−0.06	0%	27%
Mother’s university-level education	0.09	0.48	23%	−0.03	−0.04	0%	23%
Living with both parents	0.04	0.12	1%	−0.04	−0.07	0%	2%
Knowledge of family income	0.18	0.3	9%	−0.25	−0.39	15%	24%
Reduction in family income	0.28	0.4	16%	−0.13	−0.15	2%	18%
Receive pocket money	−0.04	−0.2	4%	−0.25	−0.32	10%	14%
Reduction in pocket money	−0.12	−0.18	3%	0	−0.19	4%	7%
Keeping records of income/expenses	0.08	0.08	1%	−0.76	−0.8	64%	65%

Notes: For emphasis, structure coefficients above 0.40 are underlined in Table 6 (following a convention in many factor analyses). Communalities above 40 percent are also underlined to show the variables with the highest level of usefulness in the model.

presents the canonical weights (standardized coefficient) with “w” and the canonical loadings (structure correlations) with “r_s”, the squared canonical loadings with “r_s²”, and the communalities with “h²” across the two functions for each variable. Given the sample size of 3028 students, those canonical loadings are significant for interpretative purposes (Hair et al., 2019).

The loadings for Canonical Function 1 show that the significant predictor variables are primarily “excellent” performance in school (0.81), father’s university-level education (0.52), mother’s university-level education (0.48), and students’ perception of parents’ income change having made a secondary contribution to the synthetic predictor variable (|r_s| > 0.40). This conclusion is supported by the squared canonical loadings. These students’ characteristics have larger canonical weights. A slight exception is the mother’s university-level education that has a modest canonical weight but large canonical loadings. Furthermore, all those variables’ canonical loadings are positively related.

Regarding Canonical Function 1, the financial knowledge variable is the primary contributor to the predictor synth etic variable. Because the canonical loading is positive, it is positively related to all the above students’ characteristics. Thus, Canonical Function 1 is labelled as “grade in school, parents’ university-level education, perception of family’s income change, and financial knowledge of 15-year-olds students”.

By examining Canonical Function 2, the canonical loadings in Table 7 show that the only prime predictor variable of relevance is “keeping records of income and expenses” (|r_s| > 0.40. Just for reference reasons, the author states that the “knowledge of family’s income” variable almost has an effect on the synthetic predictor variable with a canonical loading of 0.39. This student’s characteristic is inversely related to this function. The dominant predictor for Canonical Function 2 is financial behaviour (0.90). Also, by checking the signs of the canonical loadings for the entire function, a student’s characteristic of keeping records of income and expenses displays a greater financial behaviour. Given the nature of these variables, this function is labelled as “keeping records on income and expenses and financial behaviour of 15-year-olds students”.

Overall, the canonical correlation analysis serves as a strong alternative multivariate test of existing relationships between students’ characteristics and their effect on students’ financial literacy. By examining the size and the signs of the canonical loadings across the two functions has validated the findings of the analysis by indicating that an “excellent” grade in school, parents’ university-level of education, perception of family’s income reduction, and keeping records of income and expenses are the significant predictors of financial literacy.

6. Discussion and Concluding Remarks

Numerous studies by international, government, private, and non-government organisations have documented the importance of financial literacy (e.g., Hahn, 2014; Klapper et al., 2015; Lusardi & Mitchell, 2006, 2014; Mandell, 2008; OECD, 2014, 2017, 2020, 2024a, 2024b; OECD INFE, 2016a, 2016b). Atkinson (2018) stated that for a national approach to financial literacy to be fully coordinated, measuring the levels of financial literacy and identifying the predictors of financial literacy are initially required to identify the weak areas of financial knowledge and then build initiatives for financial education. Therefore, by identifying the predictors of financial literacy at the national level, the financial education treatment can be more targeted to benefit the population and especially the younger population. Since Greece is still forming its national financial literacy strategy, the results of the study should be taken into consideration.

This work studies the links between financial literacy and students’ characteristics by using a rich novel dataset of 3028 high-school students from all the regions of Greece. The questionnaire used was approved by the Ministry of Education and Religious Affairs that also approved the survey. The measure of financial literacy was performed with a method that the OECD has proposed (OECD, 2024a; OECD INFE, 2016a). The questions in the survey were a combination of questions from the literature (Hahn, 2014; Klapper et al., 2015; Lusardi & Mitchell, 2006, 2014; Mandell, 2008; OECD, 2014, 2017, 2020, 2024a, 2024b; OECD INFE, 2016a, 2016b) and were adjusted to meet the Greek standards, after which a factor analysis was performed.

By performing a canonical analysis, the author reached the conclusion that in a small, open economy like Greece, the predictors of financial literacy for high-school students are as follows: students’ performance at school, their parents’ education, knowing about a change in the family’s monthly income, and keeping records of income and expense.

The results declare that the levels of financial literacy in Greece are low among the youth. A student is financially literate if he or she scores 70 percent or over. Thus, 42.5 percent of the sample scored at or above the threshold, with the remaining 57.50 percent scoring below the threshold.

With respect to the regions of Greece’s financial literacy levels, the analysis shows that financial literacy is uneven across the country with a big range of outcomes. The highest levels are in the regions of Athens (49.2 percent), Central Greece (48.5 percent), and the North Aegean (43.4 percent). The results show that urban areas and their surroundings have higher acceptable levels of financial literacy than remote areas and their surroundings.

Among the important conclusions is that students need to know at least basic financial concepts before they finish their compulsory education to ensure financial prosperity throughout their life. Also, schools should embrace any initiative for financial education. Furthermore, schools should invite experts to talk to students about personal finance and financial management matters. Hence, as students indicated with their answers, high schools should participate in any available financial or economic competition. Through that process, students can become more familiar with financial terms and procedures.

Head teachers should understand the importance of financial literacy and teach lessons on personal finance courses at school. Either in the form of a standalone subject or in the form of being part of another subject. Also, they could include it as an elective subject or extracurricular activity after school hours. It is believed that students would benefit outmost from the financial knowledge they will obtain, as this knowledge is missing from the school curriculum. Financial knowledge is a lifelong asset for students’ financial well-being in adulthood. They will be able to make more solid financial decisions on either how to support themselves or on how to finance further studies. Hence, financial education could protect people both from potential financial shocks and from financial frauds. Globalization and digitalization have made available more sophisticated financial tools for the younger people, where targeted education is required for their financial resilience.

Parents should understand why financial education is important for their children, since the findings from the research show, respectively, that the father’s and mother’s education level have an impact on students’ financial literacy level. Financial educational material should be available for kids to learn effectively. The author suggests that measurement of financial literacy could be repeated after a “treatment” of a financial education programme. Therefore, future research could be performed after “treatment” initiatives of financial education.

The results of the study can be used by policy makers and researchers to design a more solid national strategy of financial literacy and develop initiatives for financial education especially for the young population.