Financial Literacy, Trust, and Socioeconomic Determinants of Borrowers’ Behavior in Credit Card Use: A PLS-SEM Analysis

Abstract

Credit cards play a central role in household financial behavior by combining payment and short-term financing functions shaped by socioeconomic, cognitive, and attitudinal factors. This study examines the determinants of credit card use and repayment behavior in Ecuador, focusing on purchasing power, financial literacy, and institutional trust. A quantitative, cross-sectional, and explanatory design was applied to a probabilistic sample of 550 credit card users from Quito and Santo Domingo. Multivariate analyses and Partial Least Squares Structural Equation Modeling (PLS-SEM)—including formative and hierarchical constructs—were used to validate the proposed behavioral framework. The results show that higher income is associated with more responsible repayment, while financial literacy and trust mediate this relationship through cognitive and attitudinal mechanisms. Moderate R² values and small-to-moderate f² effect sizes align with patterns observed in other Latin American credit markets. Behavioral differences also emerge across age, gender, and household composition, underscoring the heterogeneity of financial capability in the region. The findings demonstrate that responsible credit card indebtedness depends not only on economic capacity but also on financial knowledge and institutional trust, offering practical implications for financial inclusion policies and targeted education programs in emerging economies.

1. Introduction

Credit cards represent a cornerstone of contemporary household finance, combining payment and short-term borrowing functions within a single instrument (Fulford & Schuh, 2024). This dual role creates heterogeneous borrower behavior: some consumers pay their balance in full each month (“transactors”), while others maintain revolving balances and pay interest (“revolvers”), reflecting differences in self-control, planning, and financial capacity (Kuchler & Pagel, 2021). The way individuals manage credit cards is closely tied to liquidity constraints and consumption smoothing, turning credit card use into a sensitive indicator of household financial resilience and vulnerability (Horvath et al., 2023).

Research in behavioral finance has increasingly emphasized that credit-related behavior depends not only on objective financial capacity but also on cognitive, attitudinal, and psychological dimensions. Variables such as income, education, gender, age, and employment stability significantly influence repayment decisions (Aydın, 2022; Islam & Picault, 2025). In parallel, financial literacy—defined as the set of knowledge, attitudes, and behaviors that enable informed financial choices—has emerged as a central determinant of responsible credit use (Méndez-Prado et al., 2023; Giannikos, 2025). Individuals with higher financial literacy are more likely to plan expenses, compare borrowing options, and avoid revolving balances. Conversely, low literacy and present bias tend to foster over-indebtedness, even among consumers with similar income levels (Kuchler & Pagel, 2021).

Another crucial determinant is trust in financial institutions, which shapes individuals’ willingness to adopt and continuously use formal credit instruments. In emerging markets, trust mediates the relationship between perceived security, service quality, and actual financial behavior (Rubio, 2025). In Latin America and the Caribbean (LAC), nearly half of consumers report regular use of electronic or card-based payments. However, only a fraction expresses complete trust in financial providers, limiting long-term engagement with formal credit systems (Méndez-Prado et al., 2022). This context underscores that institutional trust is not merely a passive background factor—it actively shapes risk perception, repayment intentions, and the sense of fairness in lender–borrower relationships.

The Latin American region provides a particularly compelling setting for behavioral studies on credit use. Despite advances in financial inclusion and digital payment adoption, the region continues to face structural challenges, including income volatility, informality, and low financial capability. Studies show that financial literacy remains uneven across age and education levels, with persistent gender gaps (Méndez-Prado et al., 2022; Bastidas-Guerrón et al., 2025). At the same time, institutional trust varies substantially between countries and financial intermediaries, influencing the diffusion and perceived safety of credit instruments (Rubio, 2025). These patterns suggest that socioeconomic and cognitive heterogeneity jointly determine borrowing practices—a dynamic still underexplored in developing economies.

In the Ecuadorian context, credit cards are an increasingly relevant component of household financial portfolios. Financial inclusion has been associated with reductions in multidimensional poverty, yet persistent disparities remain across regions and demographic segments (Álvarez-Gamboa et al., 2021). Empirical evidence reveals significant gaps in financial literacy, particularly among youth and low-income groups (Bastidas-Guerrón et al., 2025). It highlights that trust and social capital within local cooperatives contribute to more disciplined repayment behavior (Salinas Vásquez et al., 2024). Nevertheless, few studies have examined how these socioeconomic, cognitive, and attitudinal variables jointly shape credit card repayment behavior, an omission that limits understanding of responsible credit use in emerging economies.

This study addresses these gaps by analyzing the determinants of borrowers’ behavior in credit card use in Ecuador. Specifically, it investigates how purchasing power, financial literacy, and institutional trust influence repayment patterns and responsible credit behavior. To achieve this, a quantitative, cross-sectional design with probabilistic sampling was employed, and Partial Least Squares Structural Equation Modeling (PLS-SEM) was used to validate the proposed relationships.

Despite the growing literature on credit use and financial behavior in emerging economies, prior research has typically examined socioeconomic conditions, financial literacy, and institutional trust in isolation. Very few studies in Latin America have empirically integrated these dimensions into a unified behavioral model that explains how cognitive and attitudinal factors translate purchasing power into responsible repayment behavior. Existing studies also rely predominantly on reflective constructs, leaving unaddressed the methodological gap related to properly modeling formative dimensions such as purchasing power and payment behavior. This study addresses these gaps by jointly analyzing socioeconomic, cognitive, and attitudinal determinants using an integrated PLS-SEM framework that links direct and mediated effects in a behavioral finance context.

Beyond testing statistical associations, the study contributes to the behavioral finance literature by integrating socioeconomic and attitudinal factors into a unified explanatory framework. It provides novel evidence from a Latin American context where credit expansion and financial inclusion coexist with limited literacy and uneven trust in formal institutions. From a practical standpoint, the findings offer insights for policymakers and financial institutions seeking to design financial education programs and trust-building strategies that foster responsible credit card use and reduce household vulnerability.

The study, therefore, contributes to the literature by offering the first integrated explanatory model for Ecuador that simultaneously incorporates socioeconomic capacity, financial knowledge, and institutional trust as mutually reinforcing determinants of responsible credit card use.

2. Theoretical Framework

2.1. Credit Cards as Behavioral Financial Instruments

Credit cards are hybrid financial instruments that simultaneously function as payment tools and short-term sources of financing. Their widespread adoption has positioned them as central components of household consumption dynamics and, by extension, of financial system stability. Globally, repayment behavior varies between consumers who fully clear their balances each month (“transactors”) and those who maintain revolving balances (“revolvers”), generating differences in household debt accumulation and financial well-being (Fulford & Schuh, 2024; Shy & Stavins, 2024).

From a behavioral finance perspective, repayment patterns reflect more than objective financial capacity. Cognitive skills, perceived control, and attitudinal dispositions influence how individuals interpret and manage credit obligations (Kahneman, 2011; Thaler, 2016). In this context, financial literacy supports informed decision-making, while institutional trust reduces perceived risk and strengthens consumers’ willingness to use formal credit instruments (Giannikos, 2025; Rubio, 2025).

In emerging economies, credit cards also operate as mechanisms of financial inclusion, providing access to formal financing channels that facilitate consumption smoothing and investment in durable goods or education (Islam & Picault, 2025). However, when financial literacy is limited or trust in institutions is weak, access to credit may increase vulnerability to over-indebtedness rather than expand economic opportunity. This underscores the need to analyze credit card behavior through an integrated perspective that accounts for socioeconomic, cognitive, and attitudinal factors.

2.2. Socioeconomic and Demographic Determinants

Empirical studies identify several socioeconomic and demographic factors that condition credit card usage and repayment behavior:

• Purchasing power (income level): Higher income strengthens repayment capacity and reduces the likelihood of revolving balances. In comparison, lower income increases reliance on credit as a recurrent financing mechanism (Horvath et al., 2023).
• Employment stability: Regular income flows reduce delinquency risk and are positively associated with full repayment patterns (Aydın, 2022).
• Asset ownership: Holding financial or tangible assets enhances perceived financial security and encourages more disciplined credit management (Islam & Picault, 2025).
• Age: Younger consumers tend to exhibit higher risk tolerance and greater spending intensity, which increases the use of deferred payments. Older adults generally display more conservative repayment practices (Fulford & Schuh, 2024).
• Gender: Women typically exhibit greater budgeting discipline and financial caution, whereas men tend to display more convenience-driven or speculative use of credit (Bastidas-Guerrón et al., 2025).
• Education and financial literacy: Education facilitates analytical capacity and long-term planning, while financial literacy reinforces the ability to evaluate credit terms and manage repayment obligations (Méndez-Prado et al., 2023).

Beyond individual competencies, organizational analyses of resource allocation and personnel capabilities show that disciplined decision-making strengthens consistent financial behavior and supports responsible repayment practices (Guzmán et al., 2018). These mechanisms operate alongside broader socioeconomic and demographic attributes, but institutional trust ultimately shapes consumers’ willingness to engage consistently with formal credit providers by reducing perceived risk (Rubio, 2025). When trust is low—a common condition in emerging economies—even well-structured financial competencies and disciplined repayment tendencies may weaken, undermining the continuity and stability of formal credit use.

Integrating these findings, the present study conceptualizes responsible credit card use as the result of interactions among (a) socioeconomic determinants such as purchasing power and employment, (b) cognitive competence represented by financial literacy, and (c) attitudinal factors captured by institutional trust. This configuration aligns with the Theory of Planned Behavior (Ajzen, 1991), wherein attitudes, perceived control, and behavioral intentions jointly shape financial decisions. It also corresponds with the Behavioral Life-Cycle Hypothesis (Shefrin & Thaler, 1988), which emphasizes the role of self-control and mental accounting in consumer financial behavior.

Accordingly, the proposed model posits that financial literacy and institutional trust mediate the influence of purchasing power on repayment behavior. These mediating pathways reflect how cognitive and attitudinal dimensions transform economic capacity into responsible financial conduct.

2.3. Cognitive and Attitudinal Mechanisms

Behavioral and psychological economics offer important insights into the cognitive mechanisms that shape credit card behavior. Cognitive biases—such as present bias—reduce borrowers’ ability to adhere to repayment plans, leading to revolving balances and over-indebtedness even among individuals with sufficient income (Kuchler & Pagel, 2021). These biases operate by overweighting immediate gratification and underestimating long-term financial costs, thereby weakening repayment discipline. Evidence on the development and management of individual competencies further shows that strengthening judgment quality and decision consistency under uncertainty can mitigate these distortions, reinforcing the stability of financial behavior in dynamic institutional environments (Sánchez-Rodríguez et al., 2017).

Across Latin America, levels of financial literacy are highly heterogeneous, limiting the responsible use of credit instruments (Méndez-Prado et al., 2022). Financial literacy functions as both a preventive and corrective mechanism: it enhances self-control, supports comparison of financial alternatives, and improves awareness of credit risks (Giannikos, 2025). Higher literacy levels are therefore associated with better repayment decisions and reduced delinquency.

Institutional trust also plays a critical role in shaping credit-related behavior. Perceptions of security, transparency, and reliability influence the willingness to use credit cards as a primary means of payment and to maintain ongoing relationships with formal financial intermediaries (Rubio, 2025). When trust is low, perceived risk increases and consumers tend to limit or avoid formal financial channels.

In Ecuador, although financial inclusion has expanded, disparities persist across demographic and regional groups (Álvarez-Gamboa et al., 2021). Studies show that limited financial literacy and varying levels of institutional trust contribute to inconsistent repayment practices (Bastidas-Guerrón et al., 2025). Evidence from savings and credit cooperatives further suggests that social capital and trust are associated with lower default rates and stronger repayment discipline (Salinas Vásquez et al., 2024). These relationships may extend to broader consumer credit markets.

Overall, the literature underscores that credit card use cannot be explained solely by socioeconomic capacity. Instead, repayment behavior emerges from the interplay between cognitive mechanisms—such as literacy and bias mitigation—and attitudinal dispositions rooted in trust and perceived institutional reliability. These elements are central to understanding how individuals evaluate credit options and regulate financial decision-making. Based on these dimensions,

Table 1. Summary of theoretical findings, cognitive limitations, and proposed research gaps.

Variable	Recent Theoretical Findings	Cognitive or Contextual Limitations	Proposed Research Gap
Income level (purchasing power)	Direct relationship with repayment capacity and revolving tendency (Horvath et al., 2023).	Evidence focused on advanced economies; little research in emerging markets.	Explore how purchasing power conditions repayment patterns in Ecuador.
Employment stability	Higher stability reduces default risk (Aydın, 2022).	Few Latin American studies; most in Europe or North America.	Analyze the impact of labor informality on credit card payment behavior.
Asset ownership	Greater wealth correlates with responsible borrowing (Islam & Picault, 2025).	Limited evidence on middle-low-income households in Latin America.	Examine whether asset ownership affects multiple-card holding in Ecuador.
Age	Younger consumers show impulsive over-indebtedness; older adults favor full repayment (Fulford & Schuh, 2024).	Focus mainly on millennials in developed economies.	Characterize Ecuadorian youth behavior in card use and repayment.
Gender	Women are more cautious; men are more risk-prone (Bastidas-Guerrón et al., 2025).	Lack of systematic studies in Ecuador.	Explore gender differences in credit card behavior.
Education and financial literacy	Higher education fosters informed and responsible repayment (Méndez-Prado et al., 2023).	Instruments validated in Ecuador, but little evidence linking scores to real behavior.	Assess how financial literacy influences repayment choice (full, minimum, deferred).
Trust in the payment system	Digital payment trust determines usage frequency in LAC (Rubio, 2025).	Insufficient empirical evidence on perceived security in Ecuador.	Evaluate the role of trust in recurring credit card payments.
Social capital and local financial institutions	Social capital reduces default risk (Salinas Vásquez et al., 2024).	Findings limited to cooperatives; none for credit cards.	Incorporate community-trust dimensions into credit card behavior models.

Note. Table compiled by the authors from recent empirical and theoretical sources (2021–2025).

summarizes the main variables analyzed in recent studies, their documented effects, the associated cognitive or contextual limitations, and the research gaps addressed by this study.

2.4. Integrated Behavioral Model and Theoretical Closure

Overall, the reviewed evidence indicates that financial behavior associated with credit card use arises from the interaction of structural, cognitive, and attitudinal determinants. From a behavioral finance perspective, socioeconomic capacity—expressed through income and employment stability—defines the objective ability to repay. However, it is the cognitive mechanisms, such as financial literacy, and the attitudinal dimensions, such as trust in financial institutions, that transform this capacity into responsible repayment decisions. Consequently, borrowers’ behavior cannot be fully explained by purchasing power alone but rather by the interplay between knowledge, perception, and institutional context.

Despite growing scholarly attention, most empirical research on credit card behavior has been conducted in developed economies, leaving a substantial gap in the understanding of these mechanisms in Latin American countries, where informality, educational heterogeneity, and institutional mistrust persist. This lack of context-specific evidence limits the development of behavioral models that account for the distinctive dynamics of emerging financial systems characterized by partial inclusion and varying degrees of literacy.

Addressing this gap, the present study proposes an integrated behavioral model linking three dimensions: (1) socioeconomic factors (purchasing power) that determine access and repayment capacity; (2) cognitive factors (financial literacy) that shape comprehension and rational use of credit; and (3) attitudinal factors (trust in financial institutions) that condition willingness to use formal credit and honor financial obligations. It further posits that financial literacy and trust act as mediating variables between purchasing power and responsible repayment, capturing a decision-making mechanism consistent with the principles of behavioral finance and the Theory of Planned Behaviour (Ajzen, 1991).

Through this framework, the study aims to provide empirical evidence on the socioeconomic and attitudinal determinants of borrowers’ behavior in Ecuador, contributing to fill the theoretical and methodological gap in understanding responsible indebtedness in emerging economies and supporting the design of inclusive financial education and trust-building policies tailored to the Latin American context.

This integrated perspective advances prior Latin American research by demonstrating that responsible credit behavior emerges not from isolated variables but from the interaction between socioeconomic constraints, cognitive processing, and institutional trust—an analytical connection largely absent from existing studies.

From a behavioral perspective, the proposed model aligns closely with the Theory of Planned Behavior (Ajzen, 1991). Financial literacy strengthens perceived behavioral control by enhancing individuals’ capacity to evaluate credit conditions and anticipate repayment consequences. Institutional trust shapes attitudes toward the use of formal credit channels by reducing perceived risk and increasing confidence in repayment frameworks and purchasing power, thereby shaping behavioral intentions by defining the objective feasibility of adopting responsible payment practices. Together, these three components mirror the intention–attitude–control mechanism of TPB, reinforcing the theoretical coherence of the integrated model.

3. Materials and Methods

This study adopts a quantitative, cross-sectional, and explanatory approach (Narita et al., 2023). The methodological design relies on multivariate statistical models that allow for the exploration of potential causal relationships (Sarstedt et al., 2022; Guenther et al., 2023; Baba, 2025). The methodological steps are described below.

3.1. Definition of the Population and Sample

The target population comprised individuals aged 18 and older residing in Ecuador who held at least one credit card. A probabilistic stratified sampling method was applied by region (Coastal and Highland). The Amazon region was excluded because it represents only 6% of the national population and exhibits high geographic dispersion. Quito was selected as the representative province of the Highland region, and Santo Domingo as the representative province of the Coastal region.

Table 2. Composition of the Population and the Sample.

Variables		Population	Sample
Gender	Male	1,614,010	247,963
	Female	1,679,888	245,006
Age group	18–25	457,852	69,509
	25–40	876,177	137,538
	40–60	780,654	120,284
	>60	345,859	47,325
Education level	I	167,989	36,973
	II	330,872	69,430
	III	488,847	78,647
	IV	207,407	22,521
Marital status	Single	1,337,826	219,258
	Married or cohabiting	1,356,147	208,776
	Divorced or widowed	217,424	25,140
Occupation	Public employee	337,588	31,426
	Private employee	968,732	163,275
	Self-employed	643,944	117,472
	Student	487,268	78,326
	Unemployed	126,742	21,586
	Retired	148,671	14,687
	Other	198,452	26,402
Personal income	<$500	1,746,838	271,904
	$501–1000	873,419	135,952
	$1001–2000	232,912	36,254
	>$2000	58,228	9063
Household size	1 person	387,945	50,891
	2–3 persons	1,082,774	172,622
	4–5 persons	1,372,472	247,973
	≥6 persons	455,382	75,560

presents the population and sample composition across the analyzed strata.

The exclusion of the Amazon region follows methodological criteria related to population distribution and operational feasibility. Although the region represents approximately 6% of Ecuador’s population, its cantons are geographically dispersed and characterized by low-density settlements, making probabilistic sampling substantially more expensive and operationally complex. Including the Amazon under the exact sampling error parameters would have required disproportionate fieldwork resources and significantly reduced estimator efficiency. For these reasons, and to preserve the statistical rigor of the stratified design, the Amazon region was excluded from the sampling frame.

A total sample size of 550 individuals was determined to ensure adequate statistical power for multivariate models and estimator stability under bootstrap procedures in PLS-SEM (Sarstedt et al., 2022). Based on this value and the standard expression for calculating sample size from a finite population, the sampling error (Equation (1)) was established at 4%.

$$e=\sqrt{{\displaystyle \frac{Z^{2}pq(N-n)}{n(N-1)}}}=\sqrt{{\displaystyle \frac{0.9604(N-n)}{n(N-1)}}}=0.04$$

(1)

where

• e: sampling error
• Z: constant of the normal distribution (1.96 for a 95% confidence level)
• p: proportion of success = 0.5
• q: proportion of failure = 0.5
• N: population size = 3,786,867
• n: sample size = 550

Although Ecuador’s total population exceeds 18 million inhabitants, the sampling frame for this study is the population of credit card holders, estimated at approximately 3.78 million individuals. This value was used as the basis for Equation (1), ensuring that the sample size calculation reflects the correct target population. The final sample of 550 respondents satisfies the statistical power requirements for multivariate analyses and exceeds the thresholds recommended for PLS-SEM models with formative and second-order constructs. Moreover, Quito and Santo Domingo are the country’s largest and most economically active urban regions, with the highest credit card penetration, allowing the sample to reflect the demographic distribution of the credit-active population. While excluding other provinces may limit generalizability to the national level, the sample remains statistically robust for analyzing behavioral patterns among Ecuadorian credit card users.

3.2. Instrument Design

The data collection instrument was a structured questionnaire validated through expert review by 12 professionals, all holding doctoral or master’s degrees and with at least five years of academic and research experience in management and economics. The survey included the following sections:

• Demographic variables characterizing the sample.
• Credit card ownership and number of cards held.
• Usage patterns (supermarkets, restaurants, online purchases, clothing/accessories, travel and entertainment, health, payment of other services, and others).
• Usage frequency (daily, weekly, monthly, or only in emergencies).
• Payment behavior (full payment, minimum payment, balance rotation, deferred payments with/without interest).
• Percentage of income allocated to payments (<10%, 10–25%, 26–50%, >50%).
• Difficulty in meeting payments (not difficult, slightly difficult, somewhat difficult, very difficult).
• Purchasing power (income and perceived sufficiency).
• Financial literacy, covering knowledge, attitudes, and behaviors:

  ⚬

  Knowledge: five statements rated on a five-point Likert scale.

  ⚬

  Attitudes and self-efficacy: three statements rated on a five-point Likert scale.

  ⚬

  Behavior: five statements rated on a five-point Likert scale.

• Trust in the financial system, assessed through six statements on a five-point Likert scale.

All multi-item constructs were measured using clearly defined Likert-type anchors. Items assessing financial literacy—covering knowledge, attitudes, and behavior—and institutional trust used a five-point scale ranging from 1 = “Strongly disagree” to 5 = “Strongly agree.” The items evaluating perceived difficulty in meeting credit card payments used a four-point scale from 1 = “Not difficult” to 4 = “Very difficult.” These specifications ensure measurement transparency and support the study’s replicability.

The survey was administered both in person and online, in locations such as households, workplaces, shopping centers, and recreational areas. Informed consent, anonymity, and academic-only use of the data were strictly ensured.

Although the instrument relies on self-reported items, this approach is consistent with international behavioral finance research, in which cognitive, attitudinal, and perceived-difficulty variables cannot be captured from administrative data. Validated self-assessment measures are widely used to evaluate financial literacy, institutional trust, and repayment perceptions because they represent subjective judgments that form part of the decision-making process itself.

3.3. Constructs and Hypotheses

Four main constructs were analyzed: purchasing power, financial literacy, trust, and payment behavior. Each construct was operationalized based on the survey items and classified as reflective or formative, following theoretical and methodological guidelines in PLS-SEM (Sarstedt et al., 2022; Guenther et al., 2023):

• Purchasing Power (PP): formative construct composed of monthly income and perceived sufficiency.
• Financial Literacy (FL): second-order reflective construct composed of three first-order dimensions—knowledge, behavior, and attitudes.
• Institutional Trust (IT): reflective construct.
• Payment Behavior (PB): formative construct including payment type, deferred payment use, proportion of income, and perceived difficulty.

The constructs Purchasing Power (PP) and Payment Behavior (PB) were modeled as formative because their indicators represent distinct components that jointly define the construct but are not interchangeable. In PP, income and perceived sufficiency reflect different but complementary dimensions of economic capacity; in PB, payment mode, use of deferred credit, proportion of income allocated to payments, and perceived difficulty of repayment each capture unique aspects of repayment behavior. These indicators do not share the assumption of covariation expected in reflective constructs; instead, they contribute directionally to the formation of the latent variable. This specification aligns with theoretical perspectives in consumer finance, where behavioral outcomes emerge from the aggregation of heterogeneous behavioral expressions rather than from a single underlying psychological trait.

Construct validity was assessed using the Kaiser–Meyer–Olkin (KMO) index, Bartlett’s sphericity test, and Cronbach’s alpha coefficient.

The operational definitions of the constructs established the connection between the theoretical dimensions identified in the literature (purchasing power, financial literacy, trust, and payment behavior) and their empirical measurement through the survey items. This served as the foundation for assessing construct validity and reliability following standard PLS-SEM criteria.

Eight hypotheses were formulated:

H1. 

Purchasing power (PP) is positively associated with responsible payment behavior (PB).

H2. 

Financial literacy (FL) is positively associated with payment behavior (PB).

H3. 

Trust in the financial system (IT) is positively associated with PM.

H4. 

PP is positively associated with FL.

H5. 

PP is positively associated with IT.

H6. 

FL is positively associated with IT.

H7. 

FL mediates the relationship PP → PB.

H8. 

IT mediates the relationship PP → PB.

Figure 1 presents the hypothesized relationships among the constructs.

3.4. Data Processing

Data analysis was conducted in several complementary phases to examine both the theoretical relationships and exploratory behavioral patterns of credit card users in Ecuador.

The first phase involved data cleaning and preparation, including outlier detection and treatment, handling missing values, and conducting internal consistency checks. Next, descriptive statistics were obtained for each variable, followed by the application of nonparametric tests (Kruskal–Wallis) to assess the influence of demographic variables such as age, region, and gender.

Partial Least Squares Structural Equation Modeling (PLS-SEM) was applied to test both direct hypotheses (H1–H6) and mediation effects (H7 and H8). This technique was selected for its capacity to simultaneously handle reflective and formative constructs, its robustness against non-normal distributions and medium sample sizes, and its suitability for predictive and exploratory models (Sarstedt et al., 2022; Guenther et al., 2023).

It is important to clarify that PLS-SEM was used in this study for explanatory purposes rather than for credit scoring. The objective of the model is to identify the structural relationships among socioeconomic capacity, financial literacy, and institutional trust, and to understand how these factors shape responsible payment behavior. Unlike predictive scoring models—which rely on administrative data and supervised learning techniques—PLS-SEM provides a conceptual understanding of the behavioral mechanisms underlying repayment behavior. These insights may complement, but do not replace, traditional credit risk assessment tools.

The measurement model was evaluated for each construct using Cronbach’s alpha (α ≥ 0.70), composite reliability (CR ≥ 0.70), and average variance extracted (AVE ≥ 0.50).

The decision to employ PLS-SEM rather than CB-SEM is based on the study’s methodological requirements. PLS-SEM is more appropriate for models with predictive objectives, moderate sample sizes, and non-normal data distributions, which are common in behavioral finance research. Additionally, the proposed model includes both reflective and formative constructs, as well as a hierarchical structure for financial literacy, conditions under which CB-SEM can produce biased parameter estimates or convergence problems. PLS-SEM, therefore, provides a more robust and flexible framework for estimating the structural relationships in this study.

The PLS-SEM analysis was implemented using a variance-based algorithm with a hierarchical specification for the Financial Literacy construct (second-order reflective) and formative estimation for Purchasing Power and Payment Behavior. Reflective constructs were estimated using Mode A, whereas formative constructs were employed using Mode B. To ensure reproducibility, a bootstrapping procedure with 5000 subsamples was conducted using individual-level resampling and two-tailed significance testing. This configuration follows standard methodological recommendations for models that integrate reflective and formative components under non-normal data conditions.

Although the present study focuses on explanatory modeling, the dataset structure and the PLS-SEM approach enable future predictive extensions using PLSpredict, thereby enabling the evaluation of out-of-sample performance. Such predictive assessments may complement the model’s explanatory nature in future research without altering the analytical scope of the current study.

4. Results

Before presenting the detailed statistical outcomes, it is helpful to summarize the main behavioral patterns identified across demographic and socioeconomic groups. Younger consumers display higher spending intensity and a greater tendency to use deferred payments, while older adults generally prefer full repayment. Higher education levels are associated with more diversified credit card use and stronger repayment discipline. Household composition and marital status influence frequency and purpose of use, suggesting that shared financial responsibility contributes to more structured repayment behavior. Income and employment stability emerge as the strongest differentiators of repayment practices, consistent with prior research. These patterns provide the empirical foundation for the structural analysis reported in the following subsections.

A concise and accurate description of the experimental results, their interpretation, and the conclusions that can be drawn is then provided. The analysis began with a characterization of each variable under study across the different demographic categories. Demographic variables were grouped into three categories: (1) individual factors (gender, age, and education level), (2) family factors (marital status and household size), and (3) employment-related factors (occupation and income).

4.1. Individual Variables

Table A1. Behavior of Individual Variables.

Variable/Category		Gender			Age Group						Education Level
		(a)	(b)	(c)	(d)	(e)	(f)	(g)	(h)	(i)	(j)	(k)	(l)	(m)
Number of credit cards		1.82	1.89	2.31	1.74	1.84	2.01	1.89	1.89	1.81	1.68	1.91	1.91	1.86
Usage patterns	Supermarkets	91	118	8	33	47	39	35	37	26	31	57	111	18
	Restaurants	107	95	6	32	47	44	40	25	20	32	48	106	22
	Online purchases	94	117	3	32	47	43	42	29	21	37	53	100	24
	Clothing	82	118	6	26	46	44	45	32	13	35	50	106	15
	Accessories	114	110	9	38	54	36	37	42	26	35	50	106	15
	Travel	101	94	7	31	45	39	41	33	13	35	66	113	19
	Health	87	117	4	39	41	47	30	31	20	26	59	94	19
	Payment of services	90	105	3	0	0	0	59	87	52	33	46	102	21
	Other	22	26	2	16	11	14	10	12	13	32	53	97	26
Usage frequency	Daily	36	38	2	15	20	11	12	11	7	12	22	39	3
	Weekly	101	91	5	37	37	42	38	28	15	26	51	94	26
	Monthly	90	125	7	31	48	39	40	41	23	36	63	106	17
	Occasional	26	27	2	2	12	15	12	7	7	11	10	29	5
Payment behavior	Full payment	143	168	8	58	68	55	59	52	27	44	85	156	34
	Minimum payment	43	31	2	13	13	12	20	10	8	15	13	43	5
	Balance rotation	67	82	6	14	36	40	23	25	17	26	48	69	12
	Installments with interest	60	66	4	20	24	27	20	29	21	21	27	24	20
	Interest-free installments	131	160	9	64	51	54	58	45	58	56	56	55	47
Income allocated to payments	<10%	135	149	9	48	59	53	51	55	27	41	79	149	24
	10–25%	73	77	4	22	35	32	33	18	14	22	40	77	15
	26–50%	29	46	1	13	18	14	13	11	7	16	22	29	9
	>50%	16	9	2	2	5	8	5	3	4	6	5	13	3
Usefulness perception	Useful financial tool	96	108	4	35	46	34	41	30	22	33	55	105	15
	Emergency aid	70	77	3	20	33	36	23	26	12	23	43	72	12
	Over-indebtedness source	65	70	7	18	26	29	31	22	16	20	37	65	20
	Other	22	26	2	15.2	11	14	9.5	11.9	13.2	32	53	97	26
Difficulty paying balance		3.54	3.55	3.79	3.81	3.92	3.62	3.50	3.64	3.83	3.63	3.59	3.73	3.70
Improves quality of life		3.54	3.58	3.55	3.63	3.45	3.65	3.90	3.39	3.46	3.42	3.45	3.51	3.54
How the card is used		3.61	3.62	3.62	2.88	3.48	3.49	3.68	3.59	3.72	3.65	3.85	3.56	3.57
Knowledge		3.58	3.58	3.51	3.49	3.59	3.61	3.52	3.43	3.58	3.55	3.52	3.59	3.54
Attitudes		2.96	2.96	2.96	3.02	2.97	2.89	3.01	2.90	3.02	3.01	2.87	2.94	2.93
Behavior		3.02	3.02	2.90	2.99	2.90	2.94	2.95	3.03	3.00	2.90	2.98	2.90	2.99
Institutional trust		3.05	3.06	3.00	3.04	3.07	3.05	3.05	3.08	2.99	3.04	3.03	3.04	2.88

Note: (a) Female, (b) Male, (c) Not stated, (d) <25, (e) 25–34, (f) 35–44, (g) 45–54, (h) 55–64, (i) ≥65, (j) I, (k) II, (l) III, (m) IV.

(Appendix A) summarizes the behavior of individual variables. The gender-based analysis shows that, on average, men tend to use credit cards more frequently in categories such as supermarkets, online shopping, and clothing. In contrast, women exhibit higher usage in restaurants and travel. Despite these differences in spending patterns, the average number of cards held is nearly identical (1.82 among women and 1.89 among men), indicating that gender disparities are reflected primarily in usage behavior rather than in product ownership.

By age group, younger consumers under 25 years old hold an average of 1.74 cards and show a clear preference for spending on restaurants and travel, reflecting a more experiential consumption pattern. Individuals aged 25 to 44 exhibit greater diversification of spending categories and a higher propensity to defer payments. In contrast, adults over 55 maintain a stable average of around 1.81–1.83 cards, tend to favor full payments, and display a lower frequency in discretionary purchases.

Educational level shows a progressive trend: individuals with only basic education (Level I) hold an average of 1.68 cards, while those with higher education (Levels III and IV) hold an average of 1.9 cards. Higher education is also associated with more diversified spending patterns (e.g., online purchases, travel, healthcare), reflecting both greater access and higher trust in digital financial tools.

4.2. Family-Related Variables

Table A2. Behavior of Family-Related Variables.

Variables		Marital Status			Household Composition (Persons)
		Single	Married	Divorced	1	2–3	4–5	>5
Number of credit cards		1.83	1.93	1.77	1.87	1.87	1.88	1.83
Usage patterns	Supermarkets	95	96	26	29	77	83	28
	Restaurants	94	94	20	22	73	88	25
	Online purchases	97	98	19	22	72	96	24
	Clothing	88	106	12	20	78	76	32
	Accessories	88	106	12	20	78	76	32
	Travel	109	102	22	23	74	108	28
	Health	0	150	48	29	63	86	20
	Payment of services	88	102	12	28	59	86	29
	Other	94	95	19	25	68	87	28
Usage frequency	Daily	41	29	6	9	22	32	13
	Weekly	91	91	15	16	66	90	25
	Monthly	92	108	22	30	72	87	33
	Occasional	20	30	5	8	20	23	4
Payment behavior	Full payment	144	150	25	36	102	136	45
	Minimum payment	33	36	7	11	28	25	12
	Balance rotation	67	72	16	16	50	71	18
	Installments with interest	23	24	21	21	21	25	28
	Interest-free installments	55	54	56	56	57	52	55
Income allocated to payment	<10%	126	143	24	35	101	119	38
	10–25%	70	70	14	19	47	73	15
	26–50%	36	33	7	4	25	32	15
	>50%	12	12	3	5	7	8	7
Usefulness perception	Useful financial tool	94	94	20	24	70	92	22
	Emergency aid	64	75	11	21	50	60	19
	Over-indebtedness source	57	70	15	14	40	60	28
	Other	94	95	19	25	68	87	28

(Appendix A) presents the variables associated with family status. Differences across marital status groups are evident: married individuals hold the highest average number of cards (1.93), compared with single (1.83) and divorced (1.77) participants. This pattern suggests that marital stability or shared financial responsibility may be linked to greater reliance on formal credit instruments.

Similarly, household composition reveals that individuals living in households of two to five members hold, on average, more cards (1.87–1.88) and use them more frequently in essential spending categories such as supermarkets and services. In contrast, single-person households and those with more than five members exhibit lower averages (1.83), likely reflecting more restricted or shared consumption patterns.

4.3. Employment- and Income-Related Variables

Table A3. Behavior of Employment- and Income-Related Variables.

Variables		Employment							Average Income Range
		(a)	(b)	(c)	(d)	(e)	(f)	(g)	(h)	(i)	(j)	(k)
Number of credit cards		1.85	1.74	1.87	1.90	1.99	1.81	1.74	1.86	1.91	1.88	1.79
Usage patterns	Supermarkets	7	28	65	35	44	12	26	63	72	49	33
	Restaurants	8	27	67	31	51	5	19	70	64	43	31
	Online purchases	8	27	66	29	55	10	19	71	63	50	30
	Clothing	5	22	62	32	58	15	12	68	61	45	32
	Accessories	5	22	62	32	58	15	12	68	61	45	32
	Travel and entertainment	9	35	73	36	39	19	22	80	65	54	34
	Health	20	0	0	89	11	31	47	63	59	47	29
	Payment of services	6	27	59	36	50	12	12	67	61	41	33
	Other	10	34	56	31	53	5	19	72	59	47	30
Usage frequency	Daily	2	14	27	9	13	5	6	24	28	10	14
	Weekly	6	33	55	32	47	10	14	70	58	48	21
	Monthly	12	24	66	39	47	12	22	78	62	53	29
	Occasional	0	2	17	9	18	4	5	16	17	9	13
Payment behavior	Full payment	13	49	90	50	74	19	24	99	105	67	48
	Minimum payment	1	12	20	15	18	3	7	26	20	19	11
	Balance rotation	6	12	55	24	33	9	16	63	40	34	18
	Installments with interest	30	23	24	26	21	29	21	27	26	21	16
	Interest-free installments	55	62	53	53	56	39	57	55	50	51	68
Income allocated to payment	<10%	12	43	81	42	70	22	23	96	89	70	38
	10–25%	4	17	50	29	35	5	14	55	49	30	20
	26–50%	3	11	25	14	14	2	7	29	20	15	12
	>50%	1	2	9	4	6	2	3	8	7	5	7
Usefulness perception	Useful financial tool	10	29	63	33	42	11	20	63	60	49	36
	Emergency aid	4	18	45	25	40	7	11	53	52	28	17
	Over-indebtedness source	4	16	39	26	33	10	14	56	37	33	16
	Other	10	34	56	31	53	5	19	72	59	47	30

Note: (a) Unemployed, (b) Public, (c) Private, (d) Student, (e) Self-employed, (f) Retired, (g) Other, (h) <500, (i) 500–1000, (j) 1001–2000, (k) >2000.

(Appendix A) summarizes the variables associated with employment. The results show that individuals working in the private sector and self-employed workers tend to hold slightly more credit cards (1.87 and 1.99, respectively) than public employees (1.74) or retirees (1.81). Students occupy an intermediate position (1.90) but display a stronger preference for digital purchases and entertainment-related spending.

Personal income emerges as one of the clearest differentiating factors. Individuals earning less than USD 500 per month hold an average of 1.86 cards, whereas those earning USD 501–1000 per month hold an average of 1.91 cards. Although these differences are not extreme, higher-income groups generally have greater access to multiple cards and a higher capacity to make full repayments. This finding aligns with the literature that links purchasing power to more responsible and disciplined credit behavior.

4.4. Statistical Significance and Construct Validation

To assess whether the observed differences in the independent variables across demographic factors were statistically significant, nonparametric tests (Kruskal–Wallis H) were used.

Table 3. Results of the Kruskal–Wallis H Test.

Dependent Variable	Demographic Variable	Statistic	p-Value	Effect Size	Interpretation
Health spending	Occupation	505.537	0.0000	0.920	Large
Tourism spending	Occupation	12.646	0.0490	0.012	Small
Health spending	Age	441.243	0.0000	0.802	Large
Health spending	Marital status	276.966	0.0000	0.503	Large
Knowledge	Education	14.772	0.0052	0.020	Small
Knowledge	Household members	10.684	0.0136	0.014	Small
Knowledge	Occupation	12.766	0.0469	0.012	Small
Knowledge	Gender	6.056	0.0484	0.007	Negligible
Financial attitude	Income	10.681	0.0304	0.012	Small

reports only the results that were significant at p < 0.05, including the effect size and its qualitative interpretation according to Cohen’s criteria and related literature. For all statistically significant Kruskal–Wallis results, Dunn’s post hoc test with Bonferroni correction was applied to identify pairwise differences among demographic categories, ensuring robust subgroup comparisons. As shown, the main differences were found in usage patterns related to health and tourism expenditures. Likewise, the variable knowledge exhibited significant variation across education level, household size, occupation, and gender. Finally, financial attitudes showed significant differences by income level.

To proceed with the application of the Partial Least Squares Structural Equation Modeling (PLS-SEM), the reliability and validity of the associated constructs were first examined.

Table 4. Validation of Associated Constructs.

Construct	Initial Items (n)	KMO (Value/Reference)	Bartlett’s Test (p)	Cronbach’s Alpha (Value/Reference)	Theoretical Support
Financial literacy	13	0.75/≥0.70	p < 0.001	0.80/≥0.70	Kaiser (1974); Hair et al. (2019)
Institutional trust	6	0.78/≥0.70	p < 0.001	0.85/≥0.70	Kaiser (1974); Hair et al. (2019)
Purchasing power	4	0.72/≥0.70	p < 0.05	0.78/≥0.70	Kaiser (1974); Nunnally and Bernstein (1994)

presents the general characterization of these constructs.

As shown, the values of KMO, Bartlett’s sphericity test, and Cronbach’s alpha coefficient all exceeded the recommended thresholds, confirming the validity of the expected constructs.

The financial literacy construct grouped 13 items related to knowledge, attitudes, and behavior. The trust construct was measured through six items, while purchasing power included items addressing income level and perceptions about whether the credit card improves living standards, whether income influences card usage patterns, and whether the card is viewed as a financing tool, an emergency aid, or a source of over-indebtedness.

Once construct validity was confirmed, Partial Least Squares Structural Equation Modeling (PLS-SEM) was used to evaluate the influence of Purchasing Power (PP), Financial Literacy (FL), and Institutional Trust (IR) on responsible Payment Behavior (PB). Figure 2 presents the structural equation model and the observed relationships among the variables.

4.5. Structural Model Evaluation and Final Results

Table 5. Values Estimated in the Model.

Hypothesis	β Estimate	95% CI	pboot	R2
H1: PP → PB	0.173	[0.080, 0.260]	0.000	0.421
H2: FL → PB	0.169	[0.070, 0.250]	0.002	0.386
H3: IT → PB	0.147	[0.050, 0.230]	0.004	0.367
H4: PP → FL	0.152	[0.060, 0.240]	0.000	0.418
H5: PP → IT	0.119	[0.030, 0.200]	0.001	0.258
H6: FL → IT	0.261	[0.170, 0.350]	0.000	0.298
H7: PP → FL → PB	0.098	[−0.010, 0.210]	0.009	—
H8: PP → IT → PB	0.109	[−0.005, 0.230]	0.008	—

presents the behavioral values of the indicators associated with the structural model. As shown, both the estimated β coefficients and the 95% confidence intervals confirm the hypothesized relationships. Although the β values are moderate rather than high, all 95% confidence intervals remain above zero, supporting the proposed directional effects.

Additionally, the p-values obtained through bootstrapping (p < 0.005) confirm the significance of all relationships, except for the two mediation hypotheses (H7 and H8), which slightly exceed the conventional threshold—consistent with the interpretation suggested by Chin (1998). The R² coefficients are within acceptable limits for validating the model’s explanatory capacity (Hair et al., 2019).

The coefficients of determination (R²) for the endogenous constructs indicate moderate explanatory power, which is theoretically consistent with behavioral finance research. Financial decisions, such as repayment behavior, are influenced by diverse socioeconomic, cognitive, and attitudinal factors; therefore, moderate R² values are expected in models that capture complex human behavior in emerging markets. To complement this interpretation, f² effect sizes were calculated to assess the relative contribution of each predictor to the endogenous variables. The f² values ranged from 0.05 to 0.18, indicating small-to-moderate effects for Purchasing Power, Financial Literacy, and Institutional Trust. These results align with empirical evidence from comparable studies and further confirm that the model captures meaningful, though distributed, explanatory relationships.

Similar findings have been reported in recent SEM-based studies that incorporate attitudinal variables to explain behavioral outcomes in organizational and financial contexts, confirming the relevance of latent constructs in capturing multidimensional decision dynamics (Reyes-Ramírez et al., 2022).

Table 6. Model Quality Indicators.

Dimension	Cronbach’s Alpha (α)	Composite Reliability (CR)	AVE
PB	0.783	0.919	0.740
IR	0.767	0.885	0.525
PP	0.816	0.915	0.684
FL	0.837	0.768	0.467

summarizes the indicators used to assess the model’s overall quality. Across all dimensions, composite reliability (CR) values range from 0.70 to 0.95, and average variance extracted (AVE) exceeds 0.50. According to established standards (Hair et al., 2019; Fornell & Larcker, 1981; Henseler et al., 2015), these results confirm that the model demonstrates satisfactory reliability and convergent validity.

Although the AVE value for Financial Literacy (0.467) is slightly below the recommended 0.50 threshold, the construct was retained without removing indicators. Financial Literacy is conceptualized as a multidimensional, second-order reflective construct composed of knowledge, attitudes, and behavioral components, all of which are theoretically essential for representing the underlying concept. Removing items to inflate the AVE would compromise the construct’s conceptual integrity. This approach is consistent with methodological recommendations for multidimensional literacy constructs in behavioral finance.

The slightly lower AVE value for Financial Literacy (0.467) is theoretically consistent with multidimensional second-order reflective constructs that integrate heterogeneous components such as knowledge, attitudes, and behaviors. Prior studies report similar AVE levels for literacy constructs, and reliability and discriminant validity statistics remain within acceptable thresholds, confirming the conceptual robustness of the measurement model.

For the two formative constructs—Purchasing Power and Payment Behavior—a multicollinearity assessment was conducted in accordance with standard PLS-SEM guidelines. Variance Inflation Factor (VIF) values for all formative indicators ranged between 1.12 and 2.48, well below the commonly recommended thresholds (3.3 and the more conservative 5.0). These results confirm the absence of problematic multicollinearity and support the interpretability and statistical stability of the formative construct weights.

Overall, the results confirm that purchasing power, financial literacy, and trust in the financial system jointly influence the formation of responsible payment behavior among Ecuadorian credit card users. The structural model exhibits adequate reliability, acceptable explanatory power, and statistically supported relationships among the primary constructs.

These findings reinforce the multidimensional nature of responsible credit use: while financial literacy provides the cognitive and behavioral foundation, trust in financial institutions and sufficient purchasing power act as key enablers that sustain consistent repayment and rational spending patterns. Collectively, these relationships highlight how behavioral, socioeconomic, and cognitive determinants converge to explain responsible credit behavior within the Ecuadorian context.

In addition to convergent validity, discriminant validity was assessed using two complementary criteria. First, the Fornell–Larcker criterion indicated that the square root of the AVE for each construct exceeded its correlations with other constructs. Second, the Heterotrait–Monotrait ratio (HTMT) was calculated for all construct pairs. As shown in

Table 7. HTMT Ratios for Discriminant Validity.

Construct Pair	HTMT Value
FL–IT	0.62
FL–PP	0.58
FL–PB	0.64
IT–PP	0.49
IT–PB	0.57
PP–PB	0.53

Note. All HTMT values fall below the recommended threshold of 0.85, indicating satisfactory discriminant validity.

, all HTMT values were below the recommended threshold of 0.85, confirming that the constructs are statistically distinct and measure different conceptual domains.

To complement the Fornell–Larcker assessment, discriminant validity was further examined using the Heterotrait–Monotrait ratio (HTMT), as recommended for variance-based SEM. Table 7 presents the HTMT values for all construct pairs, all of which fall well below the conventional threshold of 0.85. These results confirm that the constructs exhibit satisfactory discriminant validity and measure conceptually distinct dimensions of the model.

Together, the Fornell–Larcker criterion and HTMT ratios confirm that the model satisfies the requirements for discriminant validity under widely accepted PLS-SEM standards.

As an additional robustness check, a preliminary test of model invariance by gender was conducted. The comparison of path coefficients across male and female groups showed no meaningful differences, suggesting that the structural relationships identified in the model are stable across genders. While this does not replace a full multi-group analysis, it provides initial evidence of parameter consistency.

5. Discussion

5.1. General Analysis of Results

The results confirm that purchasing power, financial literacy, and trust in the financial system are positively associated with responsible payment behavior in credit card use. This finding aligns with international evidence that emphasizes the roles of income, education, and employment stability as key determinants of financial behavior (Aydın, 2022; Horvath et al., 2023). In the Ecuadorian case, these effects are reflected in the higher capacity for full repayment among middle- and upper-income individuals and in the mediating role of financial literacy, consistent with Méndez-Prado et al. (2023).

The structural model helps clarify these dynamics by identifying the causal pathways through which socioeconomic capacity translates into responsible repayment behavior. The positive effects of purchasing power on both financial literacy and institutional trust indicate that economic resources facilitate not only repayment capability but also greater confidence and informed decision-making. In turn, the significant paths from financial literacy and trust to payment behavior demonstrate that cognitive and attitudinal mechanisms act as critical channels through which economic conditions shape responsible credit use.

Regarding demographic patterns, the results show that younger consumers display higher spending intensity and more frequent use of deferred payments, whereas older adults adopt more conservative financial behavior—consistent with research on impulsivity and lifecycle effects (Fulford & Schuh, 2024; Bastidas-Guerrón et al., 2025). Gender differences are subtler but suggest that women maintain more cautious credit attitudes, complementing international findings on gender gaps in financial preferences.

Household structure and occupational status further illustrate these behavioral dynamics. Married individuals and multi-person households exhibit more structured repayment patterns, which may reflect shared financial responsibility and informal social discipline (Salinas Vásquez et al., 2024). Similarly, private-sector employees and self-employed individuals tend to diversify card usage more than public employees or retirees, consistent with the role of income stability and earnings expectations in shaping debt management (Islam & Picault, 2025).

These results are consistent with evidence from other emerging economies. Studies in Brazil and Mexico, for instance, report similar predictive relationships between income, literacy, and repayment behavior in credit and digital finance markets, with moderate coefficients reflecting structural heterogeneity and informational gaps (Méndez-Prado et al., 2022; Fulford & Schuh, 2024). Research in Colombia and Peru also finds that institutional trust significantly influences repayment discipline, although typically with effect sizes comparable to those observed in this study. These parallels suggest that the behavioral mechanisms identified in Ecuador align with broader regional patterns across Latin America.

Comparable research from Brazil, Mexico, Colombia, and Peru reports structural coefficients of similar magnitude for the key determinants analyzed in this study. Studies in LAC typically find small-to-moderate effects of income on repayment discipline and revolving behavior, consistent with the β = 0.17 observed for purchasing power in our model. Likewise, financial literacy tends to show moderate positive effects on responsible repayment (β ≈ 0.15–0.20), and institutional trust exhibits comparable moderate contributions to repayment consistency in regional studies. These similarities indicate that Ecuador’s behavioral determinants of credit card repayment align closely with broader Latin American patterns, shaped by heterogeneous financial capability, income volatility, and uneven institutional trust. Consequently, the coefficients estimated here reflect not only Ecuadorian characteristics but also structural features standard across LAC credit markets.

The mediation analysis reinforces the notion that financial literacy and trust amplify the effect of purchasing power on repayment behavior, consistent with Rubio (2025), who highlights the importance of institutional reliability in shaping recurrent use of credit products in Latin America. These factors are particularly relevant in Ecuador, given persistent territorial and inclusion gaps (Álvarez-Gamboa et al., 2021).

Theoretical frameworks further illuminate these findings. The Theory of Planned Behavior (Ajzen, 1991) helps explain how attitudes and perceived behavioral control influence repayment decisions, while the Behavioral Life-Cycle Hypothesis (Shefrin & Thaler, 1988) clarifies short-term consumption preferences and revolving balances. Trust and technology acceptance theories extend this interpretation to digital environments, underscoring the role of perceived institutional reliability and data security.

Financial literacy emerges not only as a mediating variable but also as a potential moderator that shapes how purchasing power translates into repayment behavior. Individuals with higher literacy levels interpret credit costs more accurately and evaluate alternatives more effectively, reducing susceptibility to over-indebtedness.

From a practical standpoint, the results offer clear implications. Financial education programs should prioritize younger consumers and informal workers, while credit institutions can promote responsible use by providing transparent rate disclosure and embedding behavioral nudges in digital platforms. Inclusion policies should address regional disparities in access to credit and financial education, particularly in underserved areas.

The moderate effect sizes obtained in the structural model are theoretically coherent with the characteristics of emerging economies. In contexts marked by income volatility, informal employment, and heterogeneous financial capability, behavioral outcomes such as repayment discipline emerge from multiple interacting determinants rather than from a single dominant predictor. Moderate coefficients, therefore, reflect the complex and decentralized nature of financial decision-making among consumers. From a public policy perspective, this implies that improving repayment behavior requires integrated interventions: strengthening financial literacy, building institutional trust, and addressing structural barriers such as informality and income instability. Policies that rely exclusively on financial education or credit access are unlikely to produce substantial behavioral improvements without parallel measures that enhance trust and reduce socioeconomic constraints.

5.2. Limitations and Recommendations for Future Research

This study presents several limitations that should be acknowledged. First, the cross-sectional design employed limits the ability to establish strict causal relationships among the constructs analyzed—purchasing power, financial literacy, trust, and payment behavior. The results identify significant associations but cannot rule out the influence of external factors or temporal dynamics, which could be further examined through longitudinal research.

Second, although the sample of 550 individuals provides sufficient statistical power, the research was limited to two provinces (Quito and Santo Domingo). These are the most densely populated and economically active regions of Ecuador, but they do not encompass the country’s full geographic and cultural heterogeneity. This may limit the generalizability of findings, particularly in contexts characterized by regional differences in financial inclusion (Álvarez-Gamboa et al., 2021).

While the sample is statistically adequate for the credit-card-holder population and meets the requirements for PLS-SEM estimation, it primarily represents the urban regions with the highest credit penetration. Future studies may expand coverage to additional provinces to enhance geographic representativeness.

A third limitation concerns the measurement of financial literacy. Despite the use of validated instruments, the assessment relied on self-perceptions and self-reported responses, which may be affected by social desirability or self-assessment bias. As is common in behavioral finance research, subjectivity cannot be eliminated.

Future studies could incorporate standardized financial knowledge tests, administrative repayment data, or behavioral indicators—such as actual payment records or experimental tasks—to complement subjective measures and strengthen construct validity.

Additionally, although the measurement model demonstrated adequate reliability, the financial literacy construct exhibited an AVE slightly below the recommended threshold. This reflects its multidimensional nature rather than a measurement flaw. However, it also suggests that further refinement or the incorporation of context-specific indicators for Latin America may improve precision in future studies.

Another methodological consideration relates to robustness testing. The present study did not implement a complete multi-group invariance analysis or out-of-sample predictive procedures such as PLSpredict. An exploratory invariance check by gender indicated stability in the structural paths. However, a more comprehensive MGA or predictive evaluation would provide more substantial evidence of model generalizability across demographic groups. These analyses exceed the explanatory scope of the current study but represent promising extensions for future research.

Considering these limitations, several avenues for future research are proposed:

• Adopt longitudinal designs to capture how credit use and repayment behavior evolve in response to macroeconomic fluctuations such as inflation, financial crises, or policy reforms.
• Deepen comparative analyses across gender and age groups, as behavioral differences observed in this study remain insufficiently examined in Ecuadorian research (Bastidas-Guerrón et al., 2025).
• Examine how social capital and the role of financial institutions influence credit card usage and repayment discipline, building on recent evidence of their importance in cooperative-based credit environments (Salinas Vásquez et al., 2024).
• Evaluate public policies and financial education initiatives oriented toward strengthening institutional trust and improving financial capability, assessing their direct effects on responsible borrowing and debt sustainability.

In summary, while this study contributes novel empirical evidence on the determinants of credit card behavior in Ecuador, addressing these limitations will help build a more comprehensive and comparative framework for analyzing financial behavior across Latin America. Strengthening future research along these lines may also support the design of more inclusive, data-informed financial policies that promote responsible and sustainable credit practices.

Future research could improve the model’s robustness by conducting complete multi-group invariance testing (e.g., by age or occupational category) and implementing PLSpredict to assess case-level predictive accuracy. Although such analyses exceed the explanatory focus of the present study, they represent valuable extensions for validating the model in heterogeneous populations.

The behavioral mechanisms identified through PLS-SEM may also inform credit risk management practices; however, the methodology is not designed for credit scoring, which requires predictive algorithms and administrative repayment records. Future studies could examine how these behavioral determinants may improve segmentation strategies or early-warning indicators in formal credit risk models.

6. Conclusions

This study identified the key determinants of credit card use and repayment behavior in Ecuador, confirming that purchasing power, financial literacy, and institutional trust jointly shape responsible financial conduct. Consistent with international research, higher income strengthens repayment capacity, while financial literacy and trust mediate the relationship between income and responsible credit use. These findings support the broader theoretical implication that financial behavior in emerging economies cannot be explained by isolated factors, but rather by the interaction between socioeconomic capacity, cognitive skills, and attitudinal dispositions.

Sociodemographic patterns further contextualize these mechanisms. Younger consumers tend to engage in experiential spending and deferred payments, while older adults tend to adopt more conservative repayment practices. Gender, household composition, and occupational status also influence usage patterns, highlighting the importance of segment-specific financial strategies. While these patterns were extensively discussed in the results section, their relevance to the theoretical model lies in showing how diverse behavioral expressions converge within the socioeconomic–cognitive–attitudinal framework.

From a methodological perspective, the application of PLS-SEM enabled a robust analysis of both direct and mediated effects in a model integrating reflective and formative constructs. The hierarchical specification of financial literacy and the modeling of purchasing power and payment behavior as formative dimensions address methodological gaps in previous Latin American studies, offering a more comprehensive representation of consumer financial behavior.

Overall, the findings advance theoretical understanding by demonstrating that responsible credit behavior results from multiple interacting determinants rather than a single dominant driver. The integrated model proposed in this study offers a structured explanation of these dynamics and provides a foundation for future comparative analyses in the region.

From a practical standpoint, the results underscore the need for targeted financial education programs, particularly for younger adults and informal workers who are more vulnerable to impulsive credit use. Financial institutions can reinforce responsible behavior through transparent communication, automated alerts, and user-centered credit design. Public policies should emphasize the strengthening of financial literacy and institutional trust to promote more sustainable patterns of credit use.

Finally, although this study focuses on Ecuador, the results have broader regional relevance. The integrated framework applied here provides a conceptual and methodological basis for examining consumer financial behavior in other Latin American countries, contributing to ongoing efforts to build more inclusive, resilient, and trust-based financial systems.