1 Introduction

Digital financial inclusion (DFI) refers to the provision of accessible and affordable formal financial services to individuals who are otherwise excluded from traditional banking systems, facilitated by digital technology. DFI uses digital tools like mobile phones and internet services to enable transactions, savings, borrowings, and insurance. The aim is to offer these services in a way that is affordable for consumers and financially viable for providers, ultimately integrating more people into the formal economy [1, 2].

Globally, DFI has attracted growing interest from academics and policymakers due to its impact on sustainable development (via inclusive growth), by bringing unbanked populations into formal financial systems through digital means [3]. Financial inclusion—particularly its digital dimension—is crucial for Saudi Arabia as it works to achieve Vision 2030’s goals of diversifying the economy, reducing oil dependency, and improving quality of life. Nationwide digital financial inclusion supports these objectives by creating a more inclusive and resilient financial system empowerment [4].

While global literature covers DFI determinants broadly (e.g., Nandru et al. [5]; Anakpo et al. [6]; Al Khub et al. [7]) research specific to Saudi Arabia remains limited. For instance, Alnemer [8] uses the Global Findex 2017 data with TAM to analyze digital banking adoption in Saudi Arabia, focusing on perceived ease, usefulness, and trust—but does not investigate disparities by gender or income. Similarly, Khan and Alhadi [9] explore fintech’s general impact on financial inclusion but do not model individual socioeconomic drivers. These studies highlight Saudi trends but do not fully address DFI disparities across demography and infrastructure. Our study fills this gap. The primary objective is to identify key drivers of DFI in Saudi Arabia, with emphasis on the 2021 wave, the first post-COVID-19 data point. By examining socio-demographic and infrastructural factors, the study offers new insights into the country’s evolving digital finance landscape.

Additionally, the study tracks the evolution of DFI across the 2011, 2014, 2017, and 2021 Global Findex waves. By revealing the main factors behind DFI, the findings guide more targeted strategies to support economic transformation—e.g., mobile-first microcredit programs for rural women or digital onboarding support for low-income youth. Ultimately, the study aims to aid policy design for enhanced digital financial services and further Vision 2030.

The remainder of the paper is organized as follows: Section two links DFI to Vision 2030 goals. Section three reviews DFI trends in Saudi Arabia from 2011–2021. Section four covers empirical literature. Section five details data and methods. Section six provides results. Section seven discusses findings and offers policy recommendations.

2 Digital financial inclusion and Saudi vision 2030

Digital financial inclusion (DFI) plays a key role in advancing Saudi Arabia’s Vision 2030 goals by improving access to formal financial services, fostering entrepreneurship, and promoting social inclusion. These outcomes support the country’s broader efforts to diversify its economy, reduce dependence on oil revenues, and improve citizens’ quality of life.

A central channel through which DFI supports Vision 2030 is by enabling small and medium-sized enterprises (SMEs) to access credit, manage payments, and grow. Programs like Kafalah provide guarantees to lenders, while digital banking and fintech platforms reduce administrative barriers and costs for small businesses. This enhances financial liquidity and operational efficiency—contributing to SME-led job creation and private-sector expansion.

DFI also improves inclusivity by extending services to underserved and remote populations through mobile banking, digital wallets, and online financial tools. In particular, online financial education platforms are becoming more important as delivery tools for improving public financial literacy and decision-making OECD [25]. These tools support financial stability and resilience by helping households budget, save, and access credit safely.

The inclusion of women and youth—especially in regions with limited mobility—is another key impact area. Digital finance allows women to access funds, launch home-based businesses, and receive payments securely. By lowering social and logistical barriers, DFI aligns with Vision 2030’s emphasis on increasing female labor force participation and empowering younger generations economically.

In parallel, the Kingdom’s push to build a strong fintech sector has helped attract international companies and investors. Government-backed initiatives like Fintech Saudi and the SAMA regulatory sandbox create a favorable environment for innovation, while digital payment systems such as Mada, SADAD, and Apple Pay improve transparency and encourage capital inflow. These developments stimulate growth in technology, data analytics, and cybersecurity—sectors crucial for long-term transformation.

DFI also enhances the efficiency of public services. By digitizing transfers, subsidies, and tax transactions, the government can streamline delivery, minimize leakages, and improve fiscal management. Initiatives like the Khazna financial literacy program further support the Vision 2030 goal of building a financially aware population.

Finally, Saudi Arabia’s ambition to become a cashless society by 2030 is directly supported by the expansion of digital payment infrastructure. Secure, interoperable, and accessible digital systems reduce the need for physical currency, enhance trust in the financial system, and support the move toward a digital-first economy.

In summary, DFI enables Saudi Arabia to pursue an inclusive, tech-enabled, and diversified economy—one of the key pillars of Vision 2030.

3 Evolution of digital financial inclusion disparities in Saudi Arabia

In this section, we analyze the evolution of digital financial inclusion (DFI) and its disparities across socio-economic groups in Saudi Arabia using cross-sectional data from the World Bank’s Global Financial Inclusion (Global Findex) database for the years 2011, 2014, 2017, and 2021.

Figures 1, 2, 3, 4, 5, 6, 7, 8 and 9 show the trends in a set of financial inclusion indicators from 2011 to 2021 across different socio-demographic groups. All figures display an upward trend, indicating progress in digital financial inclusion as reflected by various indicators. For instance, the percentage of adults in Saudi Arabia with an account at a financial institution increased from 46% in 2011 to 74% in 2021 as shown in Fig. 1.

Fig. 1
Fig. 1
Full size image

Financial institution account (% age 15 +).

Source Authors’ calculation based on Global Findex Database. These trend lines represent population-weighted point estimates from the Global Findex survey. Confidence intervals are not displayed due to limitations in data access for earlier waves, but standard errors are accounted for in regression analyses based on the 2021 wave

Fig. 2
Fig. 2
Full size image

Financial institution account by Gender.

Source Authors’ calculation based on Global Findex Database. These trend lines represent population-weighted point estimates from the Global Findex survey. Confidence intervals are not displayed due to limitations in data access for earlier waves, but standard errors are accounted for in regression analyses based on the 2021 wave

Fig. 3
Fig. 3
Full size image

Financial institution account by education level.

Source Authors’ calculation based on Global Findex Database. These trend lines represent population-weighted point estimates from the Global Findex survey. Confidence intervals are not displayed due to limitations in data access for earlier waves, but standard errors are accounted for in regression analyses based on the 2021 wave

Fig. 4
Fig. 4
Full size image

Financial institution account by income level.

Source Authors’ calculation based on Global Findex Database. These trend lines represent population-weighted point estimates from the Global Findex survey. Confidence intervals are not displayed due to limitations in data access for earlier waves, but standard errors are accounted for in regression analyses based on the 2021 wave

Fig. 5
Fig. 5
Full size image

Made or received a digital payment (% age 15 +).

Source Authors’ calculation based on Global Findex Database. These trend lines represent population-weighted point estimates from the Global Findex survey. Confidence intervals are not displayed due to limitations in data access for earlier waves, but standard errors are accounted for in regression analyses based on the 2021 wave

Fig. 6
Fig. 6
Full size image

Used a mobile phone or the internet to pay bills (% age 15 +).

Source Authors’ calculation based on Global Findex Database. These trend lines represent population-weighted point estimates from the Global Findex survey. Confidence intervals are not displayed due to limitations in data access for earlier waves, but standard errors are accounted for in regression analyses based on the 2021 wave

Fig. 7
Fig. 7
Full size image

Made a utility payment: using a mobile phone (% age 15 +).

Source Authors’ calculation based on Global Findex Database. These trend lines represent population-weighted point estimates from the Global Findex survey. Confidence intervals are not displayed due to limitations in data access for earlier waves, but standard errors are accounted for in regression analyses based on the 2021 wave

Fig. 8
Fig. 8
Full size image

Financial institution account by age.

Source Authors’ calculation based on Global Findex Database. These trend lines represent population-weighted point estimates from the Global Findex survey. Confidence intervals are not displayed due to limitations in data access for earlier waves, but standard errors are accounted for in regression analyses based on the 2021 wave

Fig. 9
Fig. 9
Full size image

Financial institution by employment status.

Source Authors’ calculation based on Global Findex Database

Figure 2 reveals a significant rise in the percentage of females with a financial institution account, from 15% in 2011 to 63% in 2021, compared to 82% of males in 2021, reflecting government efforts to promote women’s empowerment in Saudi Arabia. However, Fig. 2 also indicates a persistent gender gap in financial inclusion, as the proportion of females holding financial institution accounts remains lower than that of males.

Figures 3 and 4 highlight persistent inequalities in financial inclusion based on education and income levels. Specifically, the wealthiest 60% of Saudi adults are more financially included than the poorest 40%, though this inequality has lessened over time. In 2011, while only 33% of the poorest 40% of Saudi adults had a financial institution account, 55% of the wealthiest 60% did. By 2021, 67% of adults in the poorest 40% held an account, compared to 79% in the wealthiest 60%. This is displayed in Fig. 4.

Similarly, financial inclusion disparities based on education level are evident. In 2011, 39% of Saudi adults with a primary education or less reported having a financial institution account, compared to 50% of those with secondary education or higher. Notably, by 2021, this educational gap in financial inclusion had disappeared, as the inclusion rates for both educational groups reached parity as depicted by Fig. 3.

Figure 5 shows a substantial increase in the percentage of Saudi adults making or receiving digital payments, from 51% in 2011 to 73% in 2021.

Additionally, the proportion of adults making utility payments via mobile phones rose almost eightfold from 6% in 2014 to 43% in 2021 as shown in Fig. 6.

Figure 7 reveals growth in digital government-to-person (G2P) transfers, rising from 13% in 2014 to 28% in 2021, indicating expanding use of digital platforms for public sector disbursements.

Figure 8 illustrates a steady increase in debit card ownership, from 45% in 2011 to 67% in 2021, underscoring improved access to formal financial instruments.

Figure 9 shows that adults in Saudi Arabia have become more likely to use financial services for saving purposes, with the share saving formally increasing from 13% in 2014 to 30% in 2021.

Figures 10, 11, 12, 13, 14, 15 and 16 reveal notable disparities in DFI by employment status, education level, income level, internet access, and mobile phone ownership.

Fig. 10
Fig. 10
Full size image

Financial Inclusion by Employment Status.

Source Authors’ calculation based on 2021 Global Findex Database

Fig. 11
Fig. 11
Full size image

Financial Inclusion by Education level.

Source Authors’ calculation based on 2021 Global Findex Database

Fig. 12
Fig. 12
Full size image

Financial Inclusuion by income level.

Source Authors’ calculation based on 2021 Global Findex Database

Fig. 13
Fig. 13
Full size image

Financial inclusion by internet access.

Source Authors’ calculation based on 2021 Global Findex Database

Fig. 14
Fig. 14
Full size image

Financial inclusion by owning a mobile phone.

Source Authors’ calculation based on 2021 Global Findex Database

Fig. 15
Fig. 15
Full size image

Financial inclusion by age groups.

Source Authors’ calculation based on 2021 Global Findex Database

Fig. 16
Fig. 16
Full size image

Financial inclusion by Gender.

Source Authors’ calculation based on 2021 Global Findex Database

Figure 10 shows that employed individuals are more likely to have a financial institution account and use it for digital transactions compared to unemployed and inactive individuals.

Figure 11 confirms that adults with tertiary education consistently exhibit higher usage of digital financial services, particularly in online transfers and payments.

According to Fig. 12, individuals in the highest income quintile report greater engagement in DFI activities than those in lower-income brackets, although the gap narrowed slightly between 2017 and 2021.

Figure 13 demonstrates that people with internet access have significantly higher participation in digital transactions. For instance, in 2021, 77% of adults with internet access reported making digital payments versus only 41% without access.

Figure 14 highlights the role of mobile phone ownership in facilitating access to digital financial tools. In 2021, 69% of mobile owners used digital payment services, compared to only 28% among non-owners.

Figure 15 provides insight into the age-wise distribution of digital transactions. Those aged 25–35 remain the most active, with over 75% using digital channels in 2021.

Figure 16 further explores gender gaps in advanced digital services such as online merchant payments. While men still lead in usage, the gap has narrowed, particularly in mobile-based commerce.

The data from 2011 to 2021 demonstrate a significant upward trend in DFI across various socio-demographic groups in Saudi Arabia, with large growth in digital payments and mobile-based activities. Despite these improvements, disparities based on gender, income, and education continue to limit financial service access for significant segments of the population. While these gaps have narrowed over time, they suggest that targeted efforts could further reduce financial inclusion inequalities.

In summary, Saudi Arabia’s DFI statistics highlight that socio-demographic, geographic, and institutional factors interact complexly to influence DFI. Although substantial progress has been made in DFI overall, challenges remain, particularly concerning inequalities based on gender, rural–urban divides, and the digital gap affecting underserved populations. These observations underscore the need for a deeper econometric analysis of these disparities, which the current study aims to provide. Such analysis can guide policymakers in designing targeted interventions to address these inequalities and promote broader access to digital financial services in alignment with Saudi Vision 2030.

4 Empirical literature

A growing literature has emerged to examine the determinants of financial inclusion across a wide range of countries, employing various data sets and methodologies. This section reviews the key empirical studies on both traditional and digital financial inclusion and highlights the geographic and thematic gaps this study seeks to address.

Studies of traditional (non-digital) financial inclusion, such as those by Abel et al. [10]; Ezzahid and Elouaourti [11]; Ozili [12, 13], and Zins and Weill [14], focused on foundational determinants like education, income, gender, and access to banking infrastructure. For example, Abel et al. [10] found that education and income levels are major contributors to increasing financial inclusion in Zimbabwe. Similarly, Ezzahid and Elouaourti [11] highlighted that improvements in educational attainment and labor market participation are vital for financial inclusion in Morocco. Ozili [13] observed that individuals with at least a secondary education in Nigeria were more likely to own a bank account, hold a debit card, and engage in formal borrowing. The author also emphasized the relevance of informal financial systems, particularly for women and individuals with lower education levels. Zins and Weill [14] reported that in 37 African countries, gender, wealth, education, age, and mobile banking were significant factors. Ozili [12] similarly identified formal account ownership, financial literacy, mobile phone access, and banking infrastructure as key drivers.

As mobile banking, internet access, and digital platforms have evolved, scholars have shifted toward studying digital financial inclusion (DFI). These recent studies recognize the increasing importance of digital literacy, mobile phone usage, and regulatory environments as new drivers. Consequently, financial inclusion now includes both traditional and digital channels.

Recent studies have explored socio-demographic determinants of DFI. Nandru et al. [5] found that gender, age, education, income, and employment status influenced DFI usage in India. Ghosh and Hom Chaudhury [15] similarly concluded that wealthier, more educated men were more likely to use DFI, and noted that India’s demonetization accelerated this shift.

In terms of geographic disparities, Liu et al. [16] studied urban–rural gaps in China, finding that industrial development and education had greater impacts in cities than in rural areas. Al Khub et al. [7] examined DFI during the COVID-19 pandemic in Jordan and found significant gaps between urban and rural areas. However, Saudi Findex lacks a rural dummy; our analysis therefore cannot test urban–rural heterogeneity, which remains a topic for future research.

Two recent studies offer important context for DFI in Saudi Arabia and its broader implications. Bajwa [17] examined user behavior toward payment apps in Saudi Arabia and found that trust—primarily shaped by perceived convenience and security—plays a central role in driving customer loyalty, while regulatory compliance had a lesser effect. This underscores the importance of user-centered design and supportive policy environments in fostering digital engagement. In another study, Bajwa [18] analyzed DFI across ASEAN countries using advanced econometric models and revealed a dual impact: DFI promotes economic growth by expanding financial access but also contributes to environmental degradation due to rising energy consumption. Their findings highlight the complex trade-offs facing policymakers and emphasize the need for green digital finance strategies to ensure long-term sustainability.

Social and institutional dimensions also matter. Evans [19] found that literacy, infrastructure, and governance supported DFI in Africa, while unemployment and weak institutions hindered it. Bathula and Gupta [20] noted that education and labor participation support both traditional and digital financial inclusion, but women and poorer individuals still face digital access barriers.

Several studies emphasize mobile technology and behavioral patterns. Anakpo et al. [6] reported that mobile ownership, digital familiarity, and active banking increased DFI in India. Their study emphasized how men with higher education and income were most likely to benefit, especially in regions with robust mobile infrastructure.

Despite progress, barriers to DFS adoption persist in developing countries. Nizam and Rashidi [21] pointed to factors like financial illiteracy, trust issues, and family-based decision-making, while also noting the influence of cash preferences and infrastructure constraints. Anakpo et al. [6] similarly flagged weak digital ecosystems, regulatory bottlenecks, and cultural habits as major hurdles.

In the MENA region, studies remain limited. Berguiga and Adair [22] focused on youth, showing how education, income, gender, and job status influence financial access, especially during the pandemic.

Elouaourti and Ibourk [23] argued that fintech and ICT access are key to expanding inclusion, although digital gender gaps persist.

In conclusion, although existing research has advanced our understanding of DFI in many countries, there is limited empirical work on Saudi Arabia, especially using Findex microdata to analyze digital inclusion trends. This study aims to fill that gap by focusing specifically on Saudi Arabia’s DFI landscape.

5 Data and methods

This paper uses individual-level cross-sectional data from the 2021 Global Findex for Saudi Arabia. The Global Findex Database, administered by the World Bank, is a key global source of data on access to financial services, derived from nationally representative surveys of approximately 128,000 adults across 123 economies conducted during the COVID-19 pandemic. For Saudi Arabia, the 2021 survey includes a wide range of financial behavior and demographic indicators collected from a representative sample of 1019 Saudi nationals. The dataset is publicly accessible through the World Bank Microdata Library, and the survey methodology is detailed in Demirgüç-Kunt et al. [24].

In the literature, financial inclusion at the individual level is commonly measured using self-reported binary indicators, where respondents disclose whether they engage in specific financial activities. These self-reported measures are well-established and have been validated in numerous national- and cross-country-level studies.

5.1 Main estimation strategy

To analyze the determinants of Digital Financial Inclusion (DFI), we estimate a series of nine separate probit regression models, each corresponding to a different binary indicator of digital financial behavior. The probit model is appropriate given the binary nature of the dependent variables. The general form of the model is expressed in Eq. (1):

$$ \begin{aligned} Digital\;Financial\;Inclusion\;Indicator_{{ik}} = & \beta _{{0,k}} + \beta _{{1,k}} Female_{i} \\ & + \beta _{{2,k}} age_{i} + \beta _{{3,k}} age_{i}^{2} \\ & + \beta _{{4,k}} Income\;quantile2_{i} \\ & + \beta _{{5,k}} Income\;quantile3_{i} \\ & + \beta _{{6,k}} Income\;quantile4_{i} \\ & + \beta _{{7,k}} Income\;quantile5_{i} \\ & + \beta _{{8,k}} Tertiary\;education_{i} \\ & + \beta _{{9,k}} Employed_{i} \\ & + \beta _{{10,k}} Internet\;access_{i} \\ & + \beta _{{11,k}} Mobile\;ownership_{i} + \varepsilon _{{ik}} \\ \end{aligned} $$
(1)

Nine binary indicators derived from the 2021 Global Findex survey for Saudi Arabia capture DFI, which is our dependent variable. These indicators include: (1) Having an account at a financial institution, (2) Using a mobile phone or internet to access an account, (3) Using a mobile phone or internet to check account balance, (4) Making bill payments online, (5) Sending money to a relative or friend online, (6) Buying something online, (7) Making a digital merchant payment, (8) Paying a utility bill using a mobile phone, (9) using a mobile phone to pay for a purchase in-store. Each of these nine outcomes is coded as a binary variable, taking the value 1 if the respondent engaged in the activity, and 0 otherwise.

To address the binary nature of the dependent variables, we estimated nine separate probit models, one for each dimension of digital financial inclusion.

The analysis controls for a standard set of individual-level demographic and socioeconomic characteristics commonly used in empirical studies, including the respondent’s age, sex, and highest level of education, employment status, and household income. Specifically, male is used as the reference category for gender, and unemployed individuals serve as the baseline for employment status. For education, primary and secondary education levels are combined into a single group labeled “secondary or below,” which serves as the omitted category in the regressions. Regarding income, respondents self-report their position on a five-point income scale representing household income quintiles. Income quintile 1—the lowest income group—is used as the reference category.

Given the potential for a nonlinear relationship between age and DFI, we include the quadratic term for age. Additionally, we account for the role of digital technology by including internet access and mobile phone ownership as covariates in the model. Internet access is measured as a binary variable equal to 1 if the respondent has internet access and 0 otherwise, with ‘no internet access’ serving as the reference group. Similarly, mobile phone ownership is a binary variable equal to 1 if the respondent owns a mobile phone and 0 otherwise, with ‘no mobile phone ownership’ as the reference category.

These baseline choices are made to ensure consistent interpretation of the marginal effects of categorical variables in the regression models. Equation (1) excludes the omitted dummy categories to avoid multicollinearity.

For a comprehensive list and definitions of the variables used in the analyses, refer to Table 1. All statistical analyses and regression results are population-weighted using the sampling weights provided in the survey, ensuring that the results accurately represent the national population. All statistical analyses were performed using Stata.

Table 1 List of variables and their definition
Full size table

5.2 Robustness checks

To ensure the robustness of the probit model findings, we conduct two complementary robustness checks to be discussed below.

5.2.1 Seemingly unrelated estimation (SUEST)

Although nine separate probit models are initially estimated, the outcome variables may be correlated due to shared unobserved determinants. To account for this, we use Stata’s suest command to combine the estimation results and test for cross-equation restrictions and joint significance of variables. This procedure also allows us to verify the stability of coefficients across models while accounting for correlated errors. The SUEST robustness check helps confirm whether gender, income, education, and mobile ownership consistently influence various aspects of digital financial inclusion, even when accounting for interdependence among the dependent variables.

5.2.2 Latent DFI index via multiple correspondence analysis (MCA)

As an alternative approach, we construct a latent digital financial inclusion index using Multiple Correspondence Analysis (MCA). This method reduces the dimensionality of the nine binary DFI indicators and creates a single continuous latent score that captures an individual’s overall level of engagement in digital finance. The first MCA dimension captures over 82% of the total inertia, making it a strong summary measure. We then use this index as the dependent variable in two regression models: (a) An OLS regression to examine how demographic and digital access variables predict continuous levels of DFI. (b) A probit regression on a binary transformation of the index (equal to 1 if the individual’s DFI score is above the sample mean, and 0 otherwise) to assess predictors of high digital financial inclusion.

These robustness checks serve two key purposes. First, they confirm the stability and direction of associations observed in the main probit models. Second, they provide complementary perspectives where the SUEST accounts for correlated error structures across financial behaviors, while the latent index approach combines the multiple outcomes into a single, unified measure of digital inclusion.

6 Empirical results

6.1 Results of the main probit estimation

Table 2 presents the average marginal effects from the nine separate probit models estimating the likelihood of participation in different dimensions of digital financial inclusion (DFI). These models examine the effects of key socio-demographic characteristics—including gender, age, income, education, employment, and access to digital technologies—on the nine DFI outcomes.

Table 2 Average Marginal Effects of the Determinants of Digital Financial Inclusion
Full size table

The results reveal that gender is a significant and consistent predictor of digital financial inclusion in Saudi Arabia. Across multiple dimensions of digital financial activity, women are systematically less engaged than men—with a few notable exceptions.

Specifically, Saudi women are 12.4 percentage points less likely than men to have an account at a financial institution, highlighting a foundational gap in basic financial access. This disparity may reflect gendered barriers such as lower financial literacy, legal or cultural constraints, or limited control over personal finances.

When it comes to online bill payment, the gender gap is even more pronounced. Women are 14.6 percentage points less likely than men to pay bills online, suggesting that even among those who are banked, digital tools for managing financial obligations remain underutilized by women. This may reflect a broader digital divide, as well as household-level decision-making norms that concentrate bill management in the hands of male members.

A similar pattern is observed in the case of sending money online to friends or relatives, where women are 12 percentage points less likely than men to use digital channels. This result underscores limited engagement in digital peer-to-peer (P2P) financial transfers, potentially due to constraints in access to linked bank accounts or the necessary digital infrastructure.

In terms of digital merchant payments, a key dimension of financial technology adoption, women are also 12 percentage points less likely than men to make any digital payment. This broad category includes purchases made online or through mobile point-of-sale platforms, suggesting a general underrepresentation of women in the emerging digital consumer economy.

Interestingly, the only digital activity where women outperform men is in using mobile phones to pay for in-store purchases. In this case, women are 9.2 percentage points more likely than men to use mobile payment applications at physical retail locations. This finding suggests that mobile-based financial tools may provide a more accessible and culturally acceptable entry point for women, especially in contexts where mobility or public interaction is constrained. The simplicity, privacy, and direct control associated with mobile financial services could be reducing traditional barriers and enabling greater female participation in everyday financial transactions.

Overall, the results emphasize the dual role of gender as both a barrier and an opportunity: while women face disadvantages in formal and online financial systems, mobile-based financial tools show potential as an inclusive mechanism to close the gender gap in digital financial engagement.

Age shows mixed effects on digital financial inclusion, with evidence of both linear and nonlinear patterns. It is positively associated with checking accounts online, making purchases, and paying bills online. However, digital payment use overall decreases with age, though the positive and significant squared term reveals a U-shaped pattern. The estimated turning points further illustrate these patterns: online checking and online purchases peak around age 33, while overall digital payment usage bottoms out and begins increasing again near age 35. For mobile in-store purchases, the effect follows an inverted U-shape, peaking around age 31 before declining. These findings highlight the importance of age-sensitive digital financial strategies, as adoption patterns vary considerably across age groups.

The income level has a consistently positive effect. Individuals in the highest income quintile are 18 percentage points more likely to have a financial account, and up to 22 percentage points more likely to pay bills or transfer money online compared to those in the poorest quintile. Similar positive gradients are observed for other digital financial behaviors.

The results indicate that tertiary education is positively associated with digital financial participation, though this association is not uniformly significant across all indicators. Notably, individuals with tertiary education are 10 percentage points more likely to send money online to friends or relatives compared to those with only secondary education or less. This suggests that higher levels of education may increase confidence in using digital peer-to-peer transfer systems, possibly due to better digital literacy or greater familiarity with financial technology platforms.

Additionally, tertiary-educated individuals are 12 percentage points more likely to make online purchases, reflecting stronger engagement with e-commerce platforms and digital marketplaces. This finding aligns with the broader literature showing that education plays a key role in reducing barriers to online financial transactions—such as concerns over security, unfamiliarity with digital tools, or limited ability to evaluate online vendors.

While tertiary education is associated with positive effects on other digital financial activities, these results are not statistically significant, suggesting that education alone may not be sufficient to close gaps in access or usage for all types of digital financial services. However, the significant effects observed in online transfers and purchases underscore the importance of education as a facilitating factor in adopting more complex or trust-dependent digital financial behaviors.

Finally, access to digital infrastructure plays a pivotal role in shaping digital financial inclusion outcomes. Mobile phone ownership significantly increases the probability of owning a bank account by approximately 33 percentage points and has strong positive effects on other key behaviors, including accessing accounts online, checking balances, paying bills, and making online purchases. For instance, mobile ownership raises the likelihood of paying bills online by 56 percentage points and checking balances by 42 percentage points. These effects are statistically significant and robust across multiple dimensions of digital financial activity. In contrast, the impact of internet access is more selective but still meaningful. While it does not appear to influence account ownership, it significantly boosts the probability of accessing accounts online and sending money digitally. These findings underscore the importance of both internet connectivity and mobile technology as complementary channels for fostering digital engagement in financial services.

6.2 Results of the robustness checks

6.2.1 Seemingly unrelated estimation (SUEST)

To test the consistency and correlation structure across the nine probit models, we conducted a robustness check using the SUEST framework. This method accounts for potential correlations in unobserved factors across the separate DFI equations and enables formal joint hypothesis testing across models.

The SUEST results, presented in Table 3, confirm the key findings from the marginal effects analysis. Gender remains a strong and statistically significant predictor in key outcomes—such as account ownership, online bill payments, online money transfers, and any digital payment. For example, the coefficient on the female variable remains negative and highly significant for owning a financial account (− 0.103), online bill payments (− 0.244), online money transfers (− 0.155), and making any digital payment (− 0.09). While the SUEST results are presented as probit coefficients rather than marginal effects, the signs and significance align with the main analysis.

Table 3 Pooled probit coefficient estimates using seemingly unrelated estimation (SUEST) for digital financial inclusion outcomes
Full size table

Income continues to show a strong and positive relationship with DFI, particularly for online transactions. The highest income group (5th quintile) is significantly more likely to engage in all forms of digital payments, confirming that household wealth is a consistent driver of financial inclusion in Saudi Arabia. Education, employment status, internet access and mobile ownership show similar directional effects as in the main models, reinforcing their importance in DFI behavior.

These findings provide reassurance that the key patterns observed in the marginal effects models are not artifacts of isolated estimations, but rather represent robust patterns across interrelated financial behaviors.

6.2.2 Latent DFI index (MCA and OLS/probit)

As an additional robustness check, we constructed a latent DFI index using Multiple Correspondence Analysis (MCA) to reduce the multiple binary DFI indicators into a single composite score. The first dimension from MCA accounted for over 82% of the total inertia, making it a strong summary measure of digital financial inclusion. We then regressed this latent index on the same set of explanatory variables using OLS and also estimated a probit model on a binary indicator for high versus low DFI (defined as above or below the sample mean).

The results of the OLS regression in Table 4 show that digital financial inclusion (DFI) in Saudi Arabia is significantly shaped by gender, income, education, and digital connectivity. Women score on average 0.38 points lower on the DFI index than men, indicating a persistent gender gap. Higher income levels are strongly associated with greater DFI; for example, individuals in the top income quintile score 0.63 points higher than those in the lowest. Similarly, individuals with tertiary education have 0.38 points higher scores than those without. Digital access plays a critical role—those with internet access and mobile ownership score 1.12 and 3.79 points higher, respectively—highlighting the important role of connectivity in promoting inclusion.

Table 4 OLS estimates of latent digital financial inclusion index
Full size table

The probit regression results in Table 5 analyze the factors influencing the likelihood of an individual having a high Digital Financial Inclusion (DFI) score, defined as being above the sample mean of the latent index constructed via Multiple Correspondence Analysis (MCA). The findings confirm that gender, income, education, and age significantly shape digital financial outcomes. Specifically, women are substantially less likely than men to achieve high digital financial inclusion, with female gender reducing the probability by approximately 46 percentage points, a statistically significant effect at the 1 percent level.

Table 5 Average marginal effects from probit on high DFI Index
Full size table

Age shows a nonlinear association: while older individuals are more likely to be digitally included, this relationship weakens with age as indicated by the negative and significant squared term. Income also plays a strong and positive role. Compared to the lowest quintile, individuals in higher income brackets—particularly those in the third, fourth, and fifth quintiles—have a significantly greater likelihood of high DFI, with marginal effects ranging from 0.56 to 0.68. Similarly, having tertiary education increases the probability of high DFI by around 27 percentage points, reinforcing the idea that education enhances digital capability and confidence.

While employment status, internet access, and mobile ownership show positive coefficients, these estimates are not statistically significant in this specification, suggesting that their effects may already be captured by other covariates or may vary across subpopulations. Taken together, the results highlight that income, education, age, and gender disparities are central to understanding digital financial inclusion gaps in the Saudi context.

To sum up, both robustness checks—SUEST and the latent DFI index model—confirm the direction and significance of the main findings. They demonstrate that: gender, age, income, employment status, and mobile/internet access are consistent determinants of DFI. The digital divide along gender and income lines is structurally significant. Mobile-based financial tools may offer more inclusive pathways, particularly for women.

These complementary models enhance the validity of the results and confirm the robustness of the main conclusions drawn from the marginal effects of the probit models.

7 Discussion and conclusion

This study provides a comprehensive analysis of the key correlates of digital financial inclusion (DFI) in Saudi Arabia. While earlier waves (2011, 2014, 2017) provide context on national trends, the econometric analysis is based solely on the 2021 cross-sectional Global Findex data. The findings, supported by insights from global studies, offer useful guidance for policymakers aiming to expand DFI while addressing socio-economic disparities. These efforts will support the broader goals of Vision 2030, particularly in promoting inclusive economic development.

The analysis shows that DFI in Saudi Arabia remains uneven across socio-economic groups. Variables such as gender, income, education, employment, age, and digital access influence financial behaviors in measurable ways. The multiple probit models applied to the 2021 data reveal that men, individuals with higher income, and those with advanced education levels are more likely to engage in digital financial activities. These patterns are consistent with global evidence that links higher socio-economic status with greater access to financial tools, digital platforms, and financial knowledge.

Employment status was positively associated with digital financial use, suggesting that individuals with regular income are more likely to adopt financial technologies. However, this is a correlational—not causal—finding, as the cross-sectional nature of the data prevents firm conclusions about directionality.

One of the most prominent disparities observed is the gender gap. Women are significantly less likely to use digital financial services than men. This mirrors trends seen across the MENA region Elouaourti and Ibourk [23] and South Asia Nandru et al. [5]. Despite Saudi Arabia’s Vision 2030 reforms to empower women, this gap persists. Encouragingly, evidence from the survey suggests that women in Saudi Arabia are increasingly using mobile payment systems for everyday purchases, indicating that mobile-based solutions may offer an entry point for broader financial participation. This is supported by the recent findings of Shahen and Sharaf [26] on the role of digital payment technologies in promoting DFI.

Similarly, income is strongly correlated with DFI. Individuals in higher income brackets are more likely to use digital services. This highlights the affordability gap and underlines the need for inclusive pricing models. Education also emerged as a key driver. Individuals with higher educational attainment are better positioned to navigate digital platforms and understand financial tools, reinforcing the need for tailored financial education programs.

Access to technology, including mobile phones and internet connectivity, was another major correlate of DFI. Yet, the dataset lacks a rural/urban indicator, which limits our ability to assess geographic disparities. This is important to acknowledge, especially since rural–urban divides are a known issue in other regional studies. Future studies should include spatial indicators to examine geographic gaps more explicitly.

To assess the robustness of our findings, two additional checks were conducted. First, we employed Seemingly Unrelated Estimation (SUEST) to jointly test whether key explanatory variables had consistent effects across the nine probit models. The SUEST results confirmed that the marginal effects—particularly for gender, income, and mobile ownership—were statistically coherent across outcomes, reinforcing the reliability of our main findings. Second, we constructed a latent Digital Financial Inclusion index using Multiple Correspondence Analysis (MCA). This index, which aggregates the nine binary indicators into a single dimension, was used as a dependent variable in both OLS and binary probit regressions. The estimated effects of the explanatory variables using this latent index aligned closely with those from the individual probit models, especially regarding the significance and direction of gender, education, and technology access. Together, these robustness checks increase confidence in the internal validity and stability of the estimated relationships.

The current study is not free from limitations. First, the analysis is based on cross-sectional data from the 2021 wave of the Global Findex database. This restricts the ability to infer causality from the observed associations. Second, the absence of a rural/urban indicator in the 2021 Saudi Findex dataset prevents us from assessing potential disparities in DFI between urban and rural areas. Future research should incorporate spatial indicators to investigate these geographic differences more thoroughly. Third, the probit regressions were conducted separately for each DFI indicator, which means potential interrelationships across different outcomes, such as mobile money use and account ownership, could not be captured. Although the SUEST procedure partially addresses this by testing cross-model consistency, more comprehensive multivariate models with correlated errors could be explored in future work. Finally, several important unobserved factors, such as trust in digital platforms, social norms, and financial risk tolerance, were not included due to data limitations. These constraints highlight the need for future studies to utilize richer datasets, including longitudinal and behavioral data, to more deeply examine the evolution of DFI in Saudi Arabia.

The policy implications of the current study’s findings can be grouped according to time horizon and intervention type. In the short term, mobile-first awareness campaigns targeted at women and low-income groups could raise usage levels. Providing financial incentives such as subsidized data plans or device vouchers could help bridge immediate access gaps. Over the longer term, financial education should be embedded into national curricula to cultivate financial capability from an early age. Tailored digital training programs for unemployed individuals and informal workers can also enhance inclusion and employability. Technology-based interventions such as expanding mobile and internet infrastructure will be essential to ensure equitable access, particularly in underserved regions. In parallel, encouraging fintech development that prioritizes user-friendly design will help attract new users. Education-based approaches should also be scaled up, including public–private partnerships for digital literacy training, and leveraging community centers and local institutions to deliver in-person support.

In conclusion, this study identifies key demographic and socio-economic correlates of DFI in Saudi Arabia, including gender, income, education, and technology access. Tackling these disparities requires a coordinated strategy that promotes financial education, improves digital access, and supports vulnerable populations. By validating the findings through robust statistical techniques—including SUEST and latent index modeling—the study strengthens its empirical contributions.

A more inclusive digital financial ecosystem will help the Kingdom advance its Vision 2030 goals while ensuring that no group is left behind. Future research using panel data, geospatial variables, and richer behavioral indicators will be essential for deepening our understanding of DFI dynamics in Saudi Arabia.