[45] defines financial inclusion as “the extent to which individuals and businesses have access to and use financial services provided by the financial system, delivered responsibly and sustainably to meet their needs”. This includes access to the full range of financial products and services, such as savings, credit, insurance, and payment facilities, to enable people to manage their money and livelihoods and grow in wealth and wellbeing. It has now become an important global debate that financial inclusion can drive economic growth, reduce poverty and, in turn, generate development opportunities.

Several theoretical constructs can be applied to financial inclusion such as [5] Theory of Planned Behaviour, Sen’s (199ach; and Rogers’) Diffusion of Innov (Ajzen), individual’s attitudes towards regarding whether other people approve of using formal services, and their perceived behavioural control over whether or not to use formal services, may determine how strongly one intends to use them.

[40] Capability Approach, that people’s capacities to use financial services well must be developed by enhancing financial literacy and skills, is also relevant. And [39] Diffusion of Innovation Theory can explain how new financial technologies spread throughout societies, as mobile money and digital financial services have in developing contexts such as Northern Ghana. These theories provide a framework for thinking about the general patterns of each key financial inclusion driver.

Numerous studies in recent years have pointed to several factors that promote financial inclusion in developing countries. [46] looked at the demand-side determinants of financial inclusion in Africa and found that being a male, richer, more educated, but older, at the same time promotes financial inclusion. Analysing 37 African countries, they arrive at the conclusion that there are strong socio-economic determinants of the low level of access and usage of formal financial services in the continent.

Likewise, [36] noted that mobile money provision serves as an incentive for saving behaviour in sub-Saharan Africa, showing how technological innovations can advance financial inclusion in areas where the traditional banking infrastructure is particularly poor. In the Ghanaian context, in a study based on the Ghana Living Standards Survey 2011, [6] identified age, literacy, wealth and network as four key determinants of financial inclusion, a study that, by focusing on the Ghanaian context, offers insights about the individual and social factors that can enhance financial inclusion.

More recently, [4] focused on the impact of digital financial services on rural financial inclusion in Ghana, highlighting the specific potential for technology-based solutions to create financial inclusion where this has historically been lacking in rural areas.

Most papers also look at impediments to financial inclusion, which [19] found to be primarily lack of documentation, distance from a financial institution, cost of services and low financial literacy. All these factors tend to be more prevalent in rural areas and among the poor, the exact group we work with in Northern Ghana. One of the biggest issues of financial inclusion nowadays is the gender gap. [32] documented that in developing countries, women have less access to formal accounts and less available credit than men.

To a certain extent, financial inclusion has been improved in Ghana. Data from the Bank of Ghana as of 2020 indicated that all parts of the country have access to formal financial services, but there is regional inequality in financial inclusion. This implies that public policy intervention is needed to promote financial inclusion in deprived areas. Furthermore, in Ghana, [1] found that financial inclusion can reduce poverty. The study found that access to financial services had a significant positive effect on household welfare; therefore, financial inclusion can serve as a tool for poverty reduction and economic development.

Yet, the literature, so far, has failed to identify the determinants of financial inclusion in the This study seeks to fill the gap by providing empirical evidence of factors that influence financial inclusion in the Northern part of Ghana using the special socio-economic and cultural context of the Northern region in Ghana as a case study, and also contributes to the extant literature on factors influencing financial inclusion in the low- and middle-income countries, which can subsequently be utilised in the formulation of policy and strategy for the financial sector in Ghana and other comparable contexts.

Since the survey was cross-sectional, it used multiple data collection methods (quantitative and qualitative) to assess people’s experiences with financial inclusion in Northern Ghana. A mixed methods approach adds value in situations where the phenomenon of inquiry is complex, because it opens doors to investigate the nuances of the forces that undergird the phenomenon at hand ([17]). Quantitative data provide evidence of relationships between variables; on the other hand, qualitative data give context and responses from respondents’ lived experiences ([28]).

The study was carried out in three districts of Northern Ghana: Tamale Metropolitan, Savelugu Municipal and Kumbungu Districts. The districts were selected based on their varied urban, peri-urban, and rural settings, the main socio-economic environments in which the scourge of financial exclusion will be examined. Besides, it encompasses the stratified sampling frame [21] suggested for a homogenous population. The study population for the study was adults (18 years and above) living in the districts. This age group was selected because it is the age at which one can legally access formal financial services in Ghana ([10]).

After drawing a multi-stage sample of 400 respondents (target sample), proportionate to the population size, the researchers followed the complex survey design proposed by [43] while determining the sample size using Cochran’s formula ([16]). The confidence level was set at 95 per cent and the margin of error at 5 per cent. In total, 338 valid questionnaires were completed and retained for analysis (response rate: 84.5%); exclusions reflected non-response and incomplete or unusable questionnaires.

First, three districts were deliberately selected to represent the urban, peri-urban, and rural areas. Then, three communities were randomly selected per district from a list of communities provided by the local authorities. Next, five households were selected systematically, starting at random within each community. Finally, one adult member per household was randomly chosen using the Kish grid method ([31]). This method ensures a random selection within each household ([33]).

Primary data was collected using both structured questionnaires and semi-structured interviews. The questionnaire was drawn from previous studies on financial inclusion ([46]; [6]) and was adapted to the local context, focusing on demographic information, use of financial services, barriers, savings behaviour, etc. Before the administration of the questionnaire, a pilot study was conducted with 30 respondents to validate the questionnaire and ensure the versions were appropriate and the questions were clearly understood ([44]).

Semi-structured interviews with key informants from financial institutions, local authorities and community leaders complemented the household data, providing contextual information and triangulating the quantitative data ([38]). Before data collection, the enumerators received training on the survey instrument and ethical considerations. The questionnaires were administered through face-to-face interviews, which allowed for iterative clarification of questions and led to higher response rates ([13]).

Data analysis was conducted using Stata 16.0 due to its powerful capabilities for handling complex survey data and performing advanced statistical analyses ([20]). All descriptive statistics, including means and standard deviation, were used to summarise the sample demographics and describe the financial inclusion patterns. These descriptive analyses help to provide a picture of the data and identify initial trends ([24]).

Logistic regression analysis was used to identify the determinants of financial inclusion and informal savings behaviour. This method is suitable for binary outcome variables. When the predicted outcomes are binary, it is appropriate to estimate the odds ratio for each predictor variable ([25]). The model specification was similar to that used by [46] for the determinants of financial inclusion in Africa.

The logistic regression model was estimated as:

If p is the probability of s = 1 The logistic regression of s on a set of predictor variables is given as;

X 1 , X 2 , X 3 , ⋯ , X k and the estimates parameter values for β 0 , β 1 , ⋯ , β k via maximum likelihood method of the following equation:

where

In terms of probability, the equation is translated as

where p is the probability of being financially included, X 1 , X 2 , ⋯ X n are the independent variables, β₀ is the constant, β 1 , β 2 , ⋯ , β n are the coefficients, and ε is the error term.

Multicollinearity among independent variables was assessed using variance inflation factors (VIF), with a threshold of 10 as [23] suggested. Goodness-of-fit was evaluated using the Hosmer-Lemeshow test and the area under the ROC curve ([25]).

Qualitative data from the semi-structured interviews were analysed using thematic analysis, following the six-step process outlined by [14]. This approach allows for identifying key themes and patterns in the qualitative data, providing deeper insights into the quantitative findings.

The study adhered to ethical guidelines for social research ([27]). Informed consent was obtained from all participants, and confidentiality of personal information was ensured. This comprehensive methodology ensures a rigorous approach to investigating the determinants of financial inclusion in Northern Ghana, combining quantitative and qualitative methods to provide a nuanced understanding of the phenomenon.

Table 1 presents summary statistics for the variables used in this study. The dependent variables are measured as 0 or 1. This means the household (or individual) has no access to an account. Age was measured in years. For age as a variable, the respondent with the highest age is 99, while the lowest is 15.18 and 70 years were considered the youngest and oldest years of respondents. The table displays the standard deviation, mean, maximum, and minimum. Total observations sum up to 338.

Table 1. Descriptive statistics.

Note: Gender is coded as 1 = Female and 0 = Male.

The gender disparity was particularly pronounced in the Northeast and Savanna Regions, while being relatively smaller in the Upper East Region. These findings align with those of [46], who reported similar gender gaps in financial inclusion across several African countries.

Account ownership was also linked to education levels. Adults with tertiary education were more than twice as likely to have a bank account as those with only primary education. The educational disparity in the Savanna region was even more pronounced, with those with a tertiary education four times as likely to have an account as those with a primary education or less. Our results confirm [4] observation that education is crucial for achieving financial inclusion among rural populations in Ghana.

The second factor was age: those aged 25 - 64 had much higher formal account ownership than younger and older age groups, with the same pattern in high-income economies as in developing economies, according to the global financial inclusion survey by [19].

The logistic regression results revealed that income, age, mobile phone use, and payment method used for wages are important determinants of financial inclusion in Northern Ghana. In addition, Table 2reports logistic regression estimates for binary savings-motive outcomes (informal savings behaviour), where the dependent variables capture whether a respondent saved for (i) farm/business, (ii) old age, and (iii) education.

Second, we found positive effects of formal education across all financial inclusion measures. Respondents with secondary and tertiary education were significantly more likely to have a formal account, save formally, and use mobile money services than respondents with primary education or less, consistent with [41] for several developing countries.

Third, although the relationship between account ownership and age was linear, it was also non-linear in age itself; that is, as the likelihood of owning an account increased with age, it began to decline after one’s 40 s. This is fully consistent with the Life Cycle Hypothesis of savings and consumption introduced by [35] when applied to the case of financial inclusion.

Moreover, with respect to gender, males were somewhat more likely, by 0.19 per cent, to have a formal account, with a probability 0.19 per cent higher than that of females. This gender gap is smaller than that found in some previous studies but is similar to the findings of [22] regarding financial inclusion in China.

Table 2.Logistic regression results for savings motives (informal savings behaviour).

Dependent variable coding: Model (1) = 1 if respondent saved for farm/business, 0 otherwise; Model (2) = 1 if respondent saved for old age, 0 otherwise; Model (3) = 1 if respondent saved for education, 0 otherwise. Gender dummy: Female = 1, Male = 0. **means significant at 5%, *means significant at 10% and ***mean significant at 1%.

Model fit: The pseudo R-squared values for these logit specifications are modest (approximately 0.037 - 0.095), which is common in cross-sectional binary-choice models and should not be interpreted as variance explained in the same way as OLS R-squared. The discussion therefore emphasises coefficient signs, magnitudes, and statistical significance as evidence on determinants of savings behaviour, alongside the substantive plausibility of the estimated relationships.

The study also identified some key barriers to financial inclusion in northern Ghana, represented in Figure 1. The figure reveals that 65% of surveyed people without any formal accounts indicated that they did not have enough money as the main reason for not having any account. About 30% of these people further indicated that lack of money was the only reason they had no bank account. This finding supports the finding by [2], who stated that income is a major barrier to financial inclusion in developing countries.

Furthermore, cost was identified as one of the barriers to having a bank account, 44% indicated that the cost of keeping an account is one of the reasons why they do not have any bank account. 25% indicated that the cost of keeping an account is the only reason why they do not have any bank account with any bank. This supports the finding by [30], who argued that affordability is a major barrier to financial inclusion in many developing countries.

Third, 20% of respondents also indicated that the distance prevented them from creating a formal account. [3] also observed similar geographical barriers in their study of the Global Findex survey. Another important consideration is lack of trust in banks as a barrier to financial inclusion. Only 13% of respondents indicated this explanation for financial exclusion (Figure 1). However, [18] identified this trust deficit as a major barrier to achieving financial inclusion in developing countries. Finally, 18% of respondents cited requirements on the documentation that was needed to be produced (Figure 1). This finding is consistent with [7], who found that documentation requirements were a major barrier to financial inclusion in Sub-Saharan Africa.

Figure 1. Barriers to financial inclusion.

These results provide insights into the challenges facing financial inclusion in Northern Ghana. We can see that financial inclusion is a multifaceted problem, and that understanding and addressing it requires solutions that encompass economics, education, culture and infrastructure. The results reveal that financial inclusion in Northern Ghana can be enhanced by targeting the identified barriers and leveraging the identified determinants.

The findings revealed that age, education level of income, financial literacy and mobile money usage contribute to financial inclusion in Northern Ghana. This finding is consistent with some studies conducted in developing countries ([46]; [4]). The positive relationship between the level of education and financial inclusion highlights the need for financial literacy programmes to increase financial access ([9]).

Income also turned out to be a key factor in whether people were financially included: those with higher income were much more likely to use formal financial services. This finding is consistent with the research of [19], who identified income as a key factor in financial inclusion worldwide.

The study also found that other common barriers to financial inclusion, such as documentation, distance to an institution, and minimum balance requirements, were prevalent. These barriers are also discussed in other developing contexts ([3]; [7]).

Intriguingly, the research revealed that the use of informal savings behaviour depends on factors such as income, age, and trust in financial institutions, underscoring the nuances of the interaction between formal and informal financial systems in developing economies. ([32]).

This study has provided insights into the determinants of financial inclusion in Northern Ghana. It was found that socio-economic factors, financial literacy, and digital financial services are critical to enhancing financial inclusion in the country.

Based on the findings, some policy prescriptions for better financial inclusion are suggested. First, the government should invest in mass financial education with context-specific content, as [34] suggested. Second, access barriers to financial inclusion, particularly geographical ones, should be overcome through adopting mobile money and other digital financial services, as was successful in other African countries ([36]). Third, banks should work with regulators to make account opening easy enough while maintaining the necessary safeguards. This has been successful in other countries such as Tanzania and Kenya.

Products for low-income consumers, such as no-minimum-balance accounts, should be developed. Pushing IFFs to be more responsive to poor consumers may be unrealistic. As [12] suggest for efforts to foster inclusive finance more generally, the authorities could nonetheless require IFFs to expand financial products for low-income consumers, such as no-minimum-balance accounts. Also important is ensuring adequate consumer protection to earn the public’s trust in formal financial institutions and thereby induce a shift from informal to formal finance ([15]).

Accordingly, tailored interventions can be designed to narrow down the gender gap, including targeted financial products and services for women, as suggested by [8]. Additionally, policymakers should promote innovative technologies, such as biometric identification, to bypass documentation barriers, as successfully done in India with Aadhaar.

The study was conducted in only three districts in Northern Ghana, and research could be extended to cover more geographical areas. Longitudinal studies could also be conducted to understand the process of financial inclusion over time. Research could also explore the effect of financial inclusion on economic outcomes at the household and community level.

Future research could, for instance, examine the extent to which informal financial networks complement formal financial services, as suggested by [29]. Work could also investigate how fintech solutions could help overcome traditional barriers to financial inclusion ([37]).