Digital financial inclusion, interoperability, and vulnerable population in Ghana: analysing the threshold effect of financial literacy

Financial inclusion is central to inclusive development because it enables households to store value safely, make and receive payments, access credit, smooth consumption, and cope with unexpected shocks (Park & Mercado, 2018; Sarpong & Nketiah-Amponsah, 2022). For vulnerable populations, however, access to formal and digital financial services remains constrained by several mutually reinforcing barriers. These include irregular and low income, lack of collateral, limited formal identification, distance from financial institutions, weak digital connectivity, low trust in formal providers, limited consumer protection, and inadequate financial knowledge (Wang & Fu, 2022; Chipunza & Fanta, 2023). As a result, vulnerable households may remain dependent on informal financial arrangements that are costly, unreliable, and poorly suited to managing shocks.

Digital financial inclusion has been promoted as one way of reducing these barriers. Mobile-money-based services can lower transaction costs, reduce distance-related constraints, facilitate remittances and payments, and expand access to savings, credit, insurance, and investment opportunities (Sahay et al., 2020; Khera, Ng, Ogawa, & Sahay, 2022). In Ghana, the expansion of mobile money and the introduction of interoperability have improved the technical possibility of sending funds across networks and platforms. Yet technical interoperability does not automatically imply inclusive use. Vulnerable households may still be excluded if they lack the financial literacy, confidence, liquidity, digital skills, or trust required to use interoperable digital financial services effectively.

Interoperability is therefore central to the present study, but it must be interpreted carefully. At the system level, interoperability refers to the technical and institutional ability of different digital financial service providers, platforms, or networks to exchange value and information seamlessly. At the household level, which is the focus of this study, interoperability is observed through the use of inter-network transactions. Thus, the empirical measure used in this paper does not claim to measure whether Ghana's mobile money systems are technically connected; rather, it captures whether a respondent has actually used interoperable functionality by making a payment across another network. This distinction is important because financial literacy cannot cause system-level interoperability, but it can influence whether households understand, trust and use interoperable services.

The motivation for this paper is therefore threefold. First, although the literature establishes that financial literacy and financial inclusion matter for welfare, less is known about the behavioural channel through which financial literacy affects the use of interoperable digital financial services among vulnerable households. Second, while interoperability is often discussed as a market-level infrastructure issue, its household-level use remains uneven, especially among marginalised groups. Low use of interoperable services can increase network lock-in, restrict remittance options, raise transaction costs, reduce the convenience of digital payments and ultimately limit the welfare benefits of digital finance. Third, most existing studies focus on financial inclusion or vulnerability separately, whereas this study integrates financial literacy, interoperable service use, digital financial inclusion and multidimensional vulnerability in a single empirical framework.

This paper answers four related empirical questions. First, does financial literacy increase household-level use of interoperable digital financial services? Second, does the use of interoperable services increase digital financial inclusion? Third, does digital financial inclusion reduce household vulnerability? Fourth, do different levels of financial literacy and different dimensions of vulnerability matter for the financial inclusion of vulnerable households? These questions are examined using household-level data from informal settlements in Ghana.

The study makes four contributions. First, it clarifies the conceptual and empirical distinction between system-level interoperability and household-level use of interoperable functionality. Second, it provides evidence from Ghanaian informal settlements, a context where digital financial services are widely promoted but where vulnerability remains high. Third, it uses a simultaneous-equation framework to examine the channels linking financial literacy, interoperable service use, digital financial inclusion and vulnerability. Fourth, it complements average effects with distributional analysis using method of moments quantile regression and proposes a domain-specific vulnerability extension that allows the separate roles of socioeconomic, demographic, housing and hygiene, healthcare and epidemiological vulnerability to be assessed.

The results show that financial literacy is positively associated with the use of interoperable services, interoperable service use is positively associated with digital financial inclusion, and digital financial inclusion is negatively associated with household vulnerability. However, the coefficient sizes are modest, and the explanatory power of the vulnerability model is limited. Accordingly, the findings should be interpreted as evidence of an incremental channel rather than as proof that financial literacy or interoperability alone can eliminate vulnerability. The findings support policy interventions that combine financial education, user-friendly interoperable infrastructure, consumer protection, affordable transaction pricing and targeted support for the most vulnerable households.

The rest of the paper proceeds as follows. Section 2 reviews the related literature. Section 3 presents the data, variable measurement and empirical strategy. Section 4 discusses the empirical results and additional vulnerability-domain analysis. Section 5 concludes and presents policy implications.

Vulnerable populations face financial constraints that are broader than lack of bank accounts. They often experience unstable income, limited savings buffers, weak access to affordable credit, low insurance uptake, limited formal documentation, low-quality housing, health-related shocks, and restricted access to formal financial institutions. These constraints make households more exposed to unexpected expenses and less able to invest in productive activities (Wang & Fu, 2022; Chipunza & Fanta, 2023). Digital finance may reduce some of these constraints by lowering distance and transaction costs, but its benefits depend on usability, affordability, trust, network coverage and the ability of users to make informed financial choices.

For households in informal settlements, the constraints are especially severe because financial exclusion interacts with spatial and socioeconomic exclusion. Poor housing and hygiene conditions can increase health-related costs, demographic vulnerabilities such as old age may reduce labour-market participation, and limited healthcare access can turn temporary shocks into persistent poverty traps. Therefore, a multidimensional measure of vulnerability is more appropriate than a single income-based measure because it recognises that financial vulnerability is produced through several overlapping channels (Acharya & Porwal, 2020; Amidu, Mensah, Ahenkan, & Bawole, 2023).

Financial literacy refers to the knowledge and skills required to understand financial concepts, compare alternatives, evaluate risk and return, manage debt, plan savings, and make informed financial decisions (Agarwalla, Barua, Jacob, & Varma, 2015; Lusardi & Tufano, 2015). Higher financial literacy is generally associated with better financial management, while low literacy is associated with poor product choice, limited savings, over-indebtedness, and weak engagement with formal financial services (Arofah, Purwaningsih, & Indriayu, 2018; Behrman, Mitchell, Soo, & Bravo, 2012). In a digital environment, literacy also affects whether users can understand fees, verify recipients, avoid fraud, interpret transaction confirmations, compare service providers, and trust digital channels.

The literature suggests that financial literacy is not only a stock of knowledge but also a behavioural enabler. Individuals with higher financial literacy are more likely to adopt and use formal and digital financial services because they are better able to understand product features and assess risks (Demirgüç-Kunt, Klapper, Singer, & Van Oudheusden, 2015; Shen, Hu, & Hueng, 2018). However, literacy alone may be insufficient when infrastructure is weak or when services are unaffordable. Hence, the effect of financial literacy is expected to be stronger when digital systems are accessible, interoperable, and trusted.

Interoperability refers to the ability of different digital financial service providers, networks, platforms or systems to connect and exchange value or information in a way that allows users to transact across providers without unnecessary technical or institutional barriers (Hodapp & Hanelt, 2022; Themistocleous, Rupino da Cunha, Tabakis, & Papadaki, 2023). In mobile money systems, interoperability enables customers of one provider to send or receive funds from customers of another provider. This can reduce fragmentation, improve convenience, expand network reach and reduce the inefficiencies associated with closed-loop systems (Bourreau & Valletti, 2015; Brunnermeier, Limodio, & Spadavecchia, 2023).

Low interoperability can harm marginalised populations in several ways. It may lock users into narrow provider networks, increase the cost and difficulty of receiving remittances, force households to maintain multiple wallets or SIM cards, raise reliance on cash, limit merchant acceptance and reduce the usefulness of digital accounts. These effects are especially burdensome for low-income households because small transaction fees, travel costs and failed transactions represent a larger share of disposable income. For this reason, interoperability is not merely a technical issue; it is a distributional issue that affects the extent to which digital finance can serve vulnerable groups.

In this paper, interoperability as a system property is distinguished from household-level use of interoperable services. Ghana may have system-level interoperability, but not every household uses it. The empirical measure therefore captures observed inter-network transaction behaviour. This construct is closer to “use of interoperable functionality” than to “technical interoperability.” The expected relationship is that financial literacy increases the likelihood that households will use interoperable functionality, and that such use broadens digital financial inclusion.

Digital financial inclusion refers to access to and use of affordable formal financial services delivered through digital channels. It includes payments, savings, withdrawals, credit, insurance, investment and receipt of transfers or income through mobile or electronic accounts (Sahay et al., 2020; Khera et al., 2022). Digital financial inclusion can reduce vulnerability by improving liquidity management, enabling faster receipt of remittances, supporting business transactions, building transaction histories and providing gateways to credit and insurance.

Empirical studies generally show that digital financial inclusion is associated with poverty reduction, income smoothing and lower vulnerability, although the strength of the effect varies across contexts (Wang & He, 2020; Chen, Liang, & Yang, 2022; Yang, Wang, Wu, & Deng, 2023). The effect may be stronger for households that combine digital access with financial knowledge and trust. Conversely, digital inclusion without literacy may expose vulnerable users to fraud, over-borrowing, or inappropriate products. This suggests the need for an integrated framework in which financial literacy, interoperable service use and digital financial inclusion operate jointly rather than independently.

The reviewed literature provides strong evidence that financial literacy, interoperability and digital financial inclusion matter for welfare, but it leaves three gaps. First, many studies treat interoperability as a system-level feature and give less attention to whether vulnerable households actually use interoperable functionality. Second, the literature often examines financial inclusion and vulnerability separately, despite the likelihood that digital financial inclusion affects vulnerability through multiple channels. Third, existing studies rarely examine whether the effects differ across financial literacy thresholds and vulnerability dimensions. Based on these gaps, the study tests the following hypotheses:

This study uses household-level data containing socioeconomic, vulnerability, digital financial inclusion, and financial literacy information. Two rounds of data were collected from 101 informal settlements across 11 administrative regions of Ghana: Greater Accra, Central, Volta, Ashanti, Western, Northern, Savanna, North-East, Bono East, Bono and Ahafo. Information on the location and number of informal settlements was obtained from People's Dialogue Ghana, a non-governmental organisation that works with informal settlements. From these settlements, 2,698 households were randomly selected. The first wave of data was collected in March 2021 and the second wave was collected in 2022.

Data quality was strengthened through tablet-based data collection, field supervision and daily data checks. Tablet collection reduced entry errors by prompting enumerators when responses were inconsistent with built-in survey checks. Completed interviews were synchronised to a central data station, allowing the survey management team to review submissions and provide feedback to field teams. Where internet connectivity was weak, interviews were synchronised when enumerators returned to district capitals. Field check tables were reviewed during fieldwork, and issues were discussed with enumerators to ensure consistency and accuracy before teams proceeded to subsequent survey areas.

Digital Financial Inclusion: Digital financial services/inclusions are financial services accessed and delivered through financial technology (fintech). We employ mobile phone as a device. That is mobile phone accounts being used for (1) savings, (2) making withdrawals, (3) usage of account for payments, (4) paying taxes, (5) receipt from business/government, (6) lending, (7) borrowing, (8) making investment, and (9) taking out insurance. We constructed an index, 1–9 with 9 being a person with more usage of financial technology and inclusion. We measure interoperability as a dichotomous variable that takes the value of 1 when the respondent is to have made a payment across another network and 0 otherwise. Vulnerable population: We employed the five-dimensions vulnerability index questionnaire by Acharya and Porwal (2020). The dimension are: (1) socioeconomic vulnerability of populations was measured with four indicator variables: attaining some level of education, employment status of respondents, monthly household income and as a proxy for poverty is the proportion of the population who did not have any household assets; (2) demographic vulnerability is measured by the use of the elderly population, place of residence, and how well homes are suited; (3) housing and hygiene conditions, the paper utilises four variables: number of people per room, availability of toilet facility within a household, availability of hand hygiene facility and household access to water; (4) availability of healthcare, the study uses four indicators - possessing of health insurance, access to health care, willingness to visit health facility and the number of health care facilities in a community and (5) epidemiological factors consider three main indicators: person's current health condition, person with any current chronic illness, person undergoing any long-term medical treatment. A respondent scores 1 when he/she demonstrates or finds himself or herself in a vulnerable situation and scores 0 when otherwise is observed. A total score of 18 denotes high vulnerability. Financial literacy is measured by assessing the basic knowledge of seven concepts necessary for financial decision making: (1) understanding of interest rate, (2) interest compounding, (3) risk diversification, (4) understanding of stock, (5) future value of a stock, (6) inflation, and (7) risk and return. A respondent scores 1 if he/she answers the question correctly, and 0 otherwise. The questions have possible answers. A total score of seven is an indication of high financial literacy, while a score of 0 indicates low financial literacy. Other control variable: This study uses a number of control variables that affect the relationship of interest. These controls include gender, age, employment status, region, marital status, and distance to the nearest bank. Gender is measured as two mutually exclusive and exhaustive categories: male or female. It takes the value of 1 when the respondent is male and 0 when otherwise is observed. The age is the age of the respondent. The region is the region in which the respondent resides. Distance is measured as how far the respondent is from a financial institution and takes the value of 1 when the respondent is far away from a bank, and 0 otherwise. Employment status connote whether the individual is in gainful employment. Marital status is measured as the married 1, otherwise 0, and similarly, business ownership is measured as 1; otherwise, it is 0.

Analysis of the relationship of interest follows the structural model proposed by Keeley (1990). We argue that households in various regions of Ghana who are financially literate can seamlessly make payments to networks other than their networks (i.e. interoperability), which in turn can affect their activities regarding their involvement in digital financial services. We further contend that given that their financial literacy enables households to make payments through other networks and on other digital platforms and services, these digital financial services can create economic opportunities for vulnerable populations. They can access financial resources, credit, and marketplaces to start or expand their businesses, diversify their income sources, and improve their livelihoods, thereby reducing vulnerability. These systems of equations are presented as follows:

where $Vulnerability$ represents the household vulnerability index, which is a continuous variable. $DigitalFinancialInclusion$ is an index that reflects access to digital financial inclusion, and $Interoperability$ takes the value of 1 if the respondent has made any transactions across networks other than their own network, and 0 otherwise. $X$ denotes covariates (such as gender, age, region, and marital status) that impact digital financial inclusion, interoperability, and vulnerability. and $∈it$ are the unobserved disturbances.

$SocioVul$ represents socioeconomic vulnerability of populations measured with four indicator variables: attaining some level of education, employment status of respondents, monthly household income and as a proxy for poverty is the proportion of the population who did not have any household assets; $DemoVul$ demographic vulnerability measured by the use of the elderly population, place of residence, and how well homes are suited; $HousingVul$ represents housing and hygiene conditions, measured by number of people per room, availability of toilet facility within a household, availability of hand hygiene facility and household access to water; $HealthcareVul$ represents healthcare vulnerability measured using four indicators - possessing of health insurance, access to health care, willingness to visit health facility and the number of health care facilities in a community; and $EpidemVul$ represents epidemiological vulnerability measured using three main indicators: person's current health condition, person with any current chronic illness, person undergoing any long-term medical treatment. A respondent scores 1 when he/she demonstrates or finds himself or herself in a vulnerable situation and scores 0 when otherwise is observed.

The descriptive statistics (see Table 1) show meaningful regional variation in vulnerability, digital financial inclusion, household-level use of interoperable services and financial literacy. The overall mean vulnerability index is 4.983 with a standard deviation of 1.898. The Volta Region records the highest mean vulnerability of 5.806, while the North East Region records the lowest mean vulnerability of 3.870. This suggests that vulnerability is not uniformly distributed across informal settlements in Ghana. Digital financial inclusion has an overall mean of 1.938 and a standard deviation of 0.986. Ahafo records the highest mean digital financial inclusion score of 2.537, whereas the Western Region records the lowest mean of 1.646.

Household-level use of interoperable services has an overall mean of 0.231, indicating that about 23.1% of respondents reported inter-network mobile money transactions. This statistic should be interpreted as usage of interoperable functionality, not as the technical availability of interoperability in Ghana. The North East Region records the highest mean of 0.391, while the Savanna Region records the lowest mean of 0.073. Financial literacy has an overall mean of 5.557 out of 7, suggesting that respondents answered a relatively high number of literacy questions correctly, although the ability to translate literacy into digital financial behaviour may differ across households.

The control variables also show relevant sample characteristics. Males constitute about 39.8% of the sample, the average respondent is approximately 38 years old, about 50.5% are married and about 82.4% are employed. These descriptive patterns justify the inclusion of demographic and socioeconomic controls in the empirical models because digital financial inclusion and vulnerability may be shaped by gender, age, marital status, employment and regional location.

Table 2 presents step-by-step effect of financial literacy on vulnerability of marginalised population through interoperability and digital financial inclusion. First, the coefficient for financial literacy is 0.045, and it is statistically significant at the 1% level. This indicates that an increase in financial literacy is associated with an increase in interoperability. Financial literacy refers to the knowledge and understanding of financial concepts and practices. When individuals have higher levels of financial literacy, they are better equipped to navigate the digital financial ecosystem, understand interoperability protocols, and utilize different systems effectively. Therefore, increased financial literacy positively affects the interoperability of digital financial services. Second, the coefficient for interoperability is 0.073, and it is statistically significant at the 1% level. This indicates that an increase in interoperability is associated with an increase in digital financial inclusion. Interoperability refers to the ability of different systems and technologies to seamlessly work together, allowing users to access and use digital financial services across platforms. Improved interoperability facilitates greater integration and compatibility between different financial systems, leading to enhanced accessibility and usage of digital financial services, which promotes digital financial inclusion. While the coefficients of financial literacy (0.045) and interoperability (0.073) are statistically significant at the 1% level, they appear relatively small in magnitude. This aligns with the expectation that financial behaviour is influenced by a constellation of interrelated factors, and no single variable exerts a disproportionately large effect. These modest values suggest that while financial literacy and interoperability are important, their real-world impacts are incremental and most effective when combined with supportive infrastructure and enabling environments. This reinforces the argument for multifaceted policy interventions that include digital infrastructure investment, user education, and ecosystem-wide interoperability. Notably, model (3), vulnerability does not show a statistically significant relationship with financial literacy. This implies that financial literacy does not immediately lessen the vulnerability of households. Rather, its impact seems to function indirectly by enhancing interoperability and expanding access to digital financial services. This result is in line with our structural model and supports earlier studies (e.g. Chen et al., 2022) that show how digital tools mediate the relationship between knowledge and results. Therefore, improving access to interoperable financial systems and financial competency must be the main goals of initiatives meant to lessen vulnerability. Third, the coefficient for digital financial inclusion is −0.134, and it is statistically significant at the 1% level. This suggests that an increase in digital financial inclusion is associated with a decrease in the vulnerability index. Digital financial inclusion refers to the accessibility and usage of digital financial services by individuals. When individuals have better access to a wide range of digital financial services, they can effectively manage their finances, have more options for savings and investment, and reduce their vulnerability to financial risks and challenges.

In terms of gender, age, and marital status, the coefficient for gender is 0.012, which is statistically significant at the 1% level. This indicates that being female is associated with a decrease in the Vulnerability Index. The coefficient for Age is 0.001 and is statistically significant at the 1% level. This means that an increase in age is associated with an increase in vulnerability. The coefficient for Marital Status is −0.013, and it is statistically significant at the 1% level. This finding suggests that being married is associated with a decrease in vulnerability. Regarding region and employment status as control variables, the coefficient for region is 0.052, which is statistically significant at the 1% level. This suggests that being located in a certain region is associated with an increase in interoperability, and thus, an increase in digital financial inclusion. The coefficient for Employment Status is 0.028, which is statistically significant at the 1% level. This implies that employment is associated with an increase in digital financial inclusion and a slight increase in interoperability.

In all three of the models shown in Table 2, the constant term has a comparatively large coefficient, but the R-squared values are low, especially in model (3) at 0.079. The results may still be impacted by unobserved heterogeneity, such as behavioral traits, trust in digital platforms, or access to financial education programs, even though our structural framework takes into consideration important demographic and economic factors like age, gender, employment, and regional disparities. Our selection of explanatory factors, however, aligns with previous research (e.g. Wang & Fu, 2022; Rahayu, Budiarti, Firdauas, & Onegina, 2023) that highlights the mediating functions of digital financial inclusion and interoperability in relating financial literacy to vulnerability outcomes.

To situate the findings from this study in the literature, several studies have also found that households with low levels of wealth and financial literacy require Internet-based financial products better suited to their needs (Lu, Guo, & Zhou, 2021). Additionally, it highlighted the challenges faced by Fintech in promoting digital financial inclusion, including low digital financial literacy and interoperability issues (Rahayu et al., 2023). These findings suggest a clear link between financial literacy and the need for digital financial products tailored to enhance interoperability. Furthermore, we demonstrate that digital financial inclusion (DFI) and its secondary indicators reduce rural household vulnerability to poverty, indicating the potential of DFI to mitigate vulnerability (Chen et al., 2022). This finding supports the idea that an increase in financial literacy, which can lead to greater DFI, may contribute to reducing vulnerability. Additionally, a study on digital financial inclusion and farmers' vulnerability to poverty in rural China revealed a negative and statistically significant relationship between digital financial service use and farmers' vulnerability (Wang & He, 2020). This suggests that, as digital financial inclusion increases, vulnerability decreases, further supporting the notion that an increase in financial literacy, leading to greater interoperability and DFI, can reduce vulnerability.

Next, we evaluated the level of financial literacy of households that promote financial inclusion and ameliorate the suffering of vulnerable populations. The same measurement of the variables was used, except for the measurement of financial literacy, which is based on households with basic knowledge of seven concepts necessary for making financial decisions: (1) understanding of interest rate, (2) interest compounding, (3) risk diversification, (4) understanding of stock, (5) future value of a stock, (6) inflation, and (7) risk and return. A total score of seven (7) is an indication of high financial literacy, while a score of 0 indicates low or no financial literacy. Based on this, we grouped the structure of our data into two categories: low (weak) knowledge in finance (household scoring 0–3), and high literacy in finance (household scoring 4–7). Here, we analyse the relationship of interest using the method of moments quantile regression. The MMQR technique offers several advantages over linear methods. One limitation of linear estimation techniques, which is addressed by the MMQR method, is that they do not consider the distribution of data and focus only on averages. Additionally, these linear models fail to account for unobserved heterogeneity across cross-sections in the panel data. Furthermore, ordinary quantile regression struggles with non-crossing estimates when calculating estimators for multiple percentiles, resulting in an invalid distribution of the responses (Halidu, Mohammed, & William, 2023). However, the MMQR method does not encounter these issues.

The panel quantile regression technique introduced by Koenker and Bassett (1978) estimates the conditional median or several quantiles of the response variables based on the specific values of the exogenous variables. Due to its limitations, Machado and Silva (2019) developed the MMQR approach, which considers fixed effects to measure the distributional and heterogeneous effects of multiple quantiles (Adebayo, Akadiri, Adedapo, & Usman, 2022). The MMQR model focuses on the estimation of location-scale Q_yit(τ|x_it) conditional quantiles, specified as follows:

Where Y_it represents the dependent variable, $Xit$ is independent $(α¡,β,θ,Ω)$ are the parameters to be estimated. Following a previous study (Amegavi, 2022), we further specify Equation (5) as follows:

The quantile distribution of the outcome variable is $Qy(μ|xit).ai(μ)$ = $ai+q(μ)$

shows the coefficient of scalar and $μth$ is the sample quantile. V represents a k-k vector.

This subsection employs the method of moments regression results for the impact of different levels of financial literacy on vulnerable populations and digital financial inclusion. To begin, Table 3 presents the results of the regressions that use the two levels of financial literacy (weak and high financial literacy) and controlling for residency-urban, educational level, and employment status. The results are presented in Columns 1 and 2 for vulnerability and Columns 3 and 4 for digital financial inclusion. Across different quantiles, the analysis indicates that individuals with weak/low financial literacy do not explain the vulnerability position in the selected sample. In contrast, those with average financial literacy experience a notable reduction in vulnerability across all quantiles, suggesting that basic financial knowledge plays a crucial role in enhancing resilience. These nuanced findings underscore the differential influence of financial literacy levels on vulnerability, emphasizing the importance of targeted interventions for specific segments of the population to address and mitigate their unique vulnerability. In this regard, the impact of average financial literacy on reducing vulnerability aligns with the findings of Behrman et al. (2012), who demonstrated that basic financial knowledge and skills are essential for making informed financial decisions and managing financial risks. This suggests that even a moderate level of financial literacy can contribute to enhancing resilience and reducing vulnerability across different quantiles.

In relation to digital financial inclusion, the results show that first, the coefficient for Weak Financial Literacy is consistently negative across all quantiles, indicating that weaker financial literacy is associated with lower digital financial inclusion. The effect is more pronounced at higher quantiles or higher levels of digital financial inclusion, as evidenced by the increasing magnitude of the coefficient. Second, the coefficient of high financial literacy is consistently positive across all quantiles, suggesting that higher financial literacy is associated with better digital financial inclusion. This effect is relatively stable across quantiles. These findings are consistent with those of Demirgüç-Kunt et al. (2015) highlighted the role of financial literacy in shaping individuals' financial behaviours and engagement with digital financial tools.

Regarding the control variables, urban location, education level, and employment status consistently show a positive and significant relationship with vulnerability, meaning that individuals residing in urban areas who are gainfully employed and have some form of education are more likely to experience financial vulnerability. These results are relatively stable across the quantiles. On the other hand, the coefficients of educational level and employment status on digital financial services are statistically significant across quantiles, suggesting that employment status and educational level have a consistent impact on digital financial inclusion. These analyses suggest that enhancing financial literacy, especially among individuals with financial literacy, could contribute to improved digital financial inclusion. Additionally, addressing rural/urban disparities may be important for promoting widespread digital financial inclusion.

Table 4 presents the additional domain-level analysis undertaken to determine whether different forms of vulnerability have similar implications for digital financial inclusion. The results show clear heterogeneity across vulnerability domains. Demographic vulnerability has a negative and statistically significant association with digital financial inclusion. This indicates that age-related and demographic disadvantages are important barriers to meaningful digital financial participation. Housing and hygiene vulnerability also has a negative and statistically significant coefficient. This means that poor living conditions are associated with lower use of digital financial services. This is plausible because poor housing conditions may affect phone security, charging access, network reliability and convenience of transacting through mobile channels.

Healthcare vulnerability is also negative and significant, while socioeconomic vulnerability is negative and significant. These results indicate that weak healthcare access, low income, unemployment, limited education, and lack of assets restrict households' ability to participate in digital finance. By contrast, epidemiological vulnerability has a negative but statistically insignificant coefficient. This suggest that current illness or long-term treatment does not independently explain digital financial inclusion after the other vulnerability domains and controls are included.

This study examined the relationship between financial literacy, household-level use of interoperable digital financial services, digital financial inclusion, and vulnerability among households in informal settlements in Ghana. The study was motivated by the observation that vulnerable households face multidimensional financial constraints, and that system-level digital infrastructure does not automatically translate into meaningful household use. Ghana's mobile money interoperability creates the possibility of inter-network transactions, but households must still understand, trust, and use that functionality for it to affect their digital financial inclusion.

The baseline evidence indicates that financial literacy is positively associated with household-level use of interoperable services, that use of interoperable services is positively associated with digital financial inclusion, and that digital financial inclusion is negatively associated with vulnerability. These findings support the channel argument of the paper: financial literacy does not simply reduce vulnerability directly; rather, it helps households engage more effectively with interoperable digital financial systems, which broadens digital financial inclusion and contributes incrementally to lower vulnerability.

The quantile results further show that weak financial literacy is associated with lower digital financial inclusion, whereas higher financial literacy supports broader digital financial service use. The domain-level analysis demonstrates that vulnerability is not homogeneous. Demographic vulnerability, housing and hygiene vulnerability, healthcare vulnerability and socioeconomic vulnerability significantly reduce digital financial inclusion. Epidemiological vulnerability is negative but statistically insignificant. These findings imply that the digital finance constraints of vulnerable households differ by the source of vulnerability and should not be addressed through a single generic inclusion policy.

The study has specific policy implications. First, financial literacy programmes should include practical digital transaction skills, including how to send inter-network transfers, verify recipient details, identify fees, read transaction confirmations, and avoid fraud. Second, regulators and providers should treat interoperability as a user-experience and inclusion issue, not only as a technical infrastructure issue. Interfaces should be simple, transaction fees transparent, failed transactions quickly resolved and consumer complaints accessible to low-literacy users. Third, policy interventions should be differentiated by vulnerability domain. Demographically vulnerable households may need assisted digital services and trusted agent support. Households facing housing and hygiene vulnerability require reliable agent access, connectivity support and safe transaction points. Healthcare-vulnerable households require insurance-linked digital products and health-payment options. Socioeconomically vulnerable households require low-fee products, income-support linkages, and targeted financial education. Fourth, financial inclusion policies should combine financial education, consumer protection, digital infrastructure, and social protection rather than relying on a single intervention.

The study is subject to limitations. First, the interoperability variable captures household-level use of inter-network transactions, not the technical level of interoperability in the national system. Second, the models explain only part of digital financial inclusion and vulnerability; unobserved factors such as trust, transaction frequency, agent quality, fraud exposure and digital confidence may also matter. Third, the empirical design identifies associations rather than definitive causal effects. Future research should examine the quality, frequency and value of interoperable transactions, use experimental or quasi-experimental designs where possible, and assess whether consumer protection and trust moderate the relationship between digital financial inclusion and vulnerability.