Formal and informal financial inclusion and its impact on health: empirical evidence from South Africa Authors: Lwando Mjacu, Vikela Liso Sithole, Palesa Makhetha-Kosi Published: 2026 ABSTRACT AbstractBackgroundThe ability to secure high quality healthcare and access at an affordable cost for the entire population is the principal health objective of many countries on the path to the attainment of Sustainable Development Goals (SDGs). This study addresses the critical, yet understudied, impact of formal and informal financial inclusion on health, directly supporting the objectives of SDG 3 (Good Health and Well-being).MethodsThe study employs and modifies South Africa’s National Income Dynamic Study microdata by combining all five waves to create panel data. A random effects ordered probit model is used since it is possibly the best approach for panel data on self-rated health outcomes as it takes into consideration the existence of an additional normally distributed cross-section error term which captures individual specific effects.ResultsResults for formal financial inclusion dummy variable are statistically significant and positive, suggesting that individuals with access to formal financial services such as savings and credit have a higher probability of reporting better health outcomes. However, results for informal financial inclusion dummy variable are statistically significant and negative, implying that higher reliance on informal finance like stokvel savings and mashonisa credit shifts individuals toward the lower outcomes of health.ConclusionGiven the mixed outcomes of the impact of formal and informal financial inclusion on health, this study recommends policies that could integrate informal savings and credit into the formal digital economy and address structural barriers that limit the effectiveness and efficiency of the financial system.Keywords: South Africa, Stokvels, Mashonisa, Self-reported health, National Income Dynamic Study, Random effects ordered probit model Introduction The ability to secure high quality healthcare and access at an affordable cost for the entire population is a principal health objective of many countries on the path to the attainment of Sustainable Development Goals (SDGs) [1]. The SDGs, also known as the Global Goals (GGs), were adopted by all United Nations Member States in 2015 as a universal call to action to end poverty, protect the planet and ensure that all people enjoy peace and prosperity by 2030 [2]. In alignment with these objectives, South Africa has introduced its most ambitious health reform ever undertaken to address inequality in access to health and geared towards achieving universal health coverage for the 60 million population [3]. The reform, which is being implemented through the National Health Insurance and Presidential Health Compact, is meant to provide access to affordable quality health care for all residents of South Africa. It is in line with the country’s constitution, the National Health Act, and the National Development Plan 2030. Priority three of the National Development Plan on health focuses on improving the quality of life and productive capacity of South Africans by improving life expectancy, reducing maternal and child mortality, and improving health outcomes for women, youth and persons with disabilities. Improving population health is essential for boosting economic performance and productivity while lowering budget deficits [4]. Healthier populations may catalyze economic performance by raising the value of investment in human and physical capital, both recognized drivers of sustained economic expansion [5]. In addition, healthier workers are physically and mentally more energetic and robust. They are more productive and earn higher wages and are also less likely to be absent from work because of illness or illness in their family [6]. This assertion is substantiated by several studies in both developing and developed countries [7–10]. Given the role that population health plays in economic advancement, numerous scholars and policymakers have worked to identify the factors that influence population health outcomes. The available studies have mostly made enquiry into the effect of factors such as income [11, 12]; employment [13, 14]; and education [15, 16] on health in both developing and developed countries. However, research on the impact of financial inclusion on health remains limited and still in the infant stage in literature. Financial inclusion, broadly defined as the provision of comprehensive, affordable, and reliable financial services [17], can potentially contribute to the attainment of SDG three. Financial inclusion has been at the forefront of national and international policy debates and has emerged as a tool and strategy for improving livelihoods worldwide, particularly in developing nations [4]. The empirical literature suggests that financial inclusion significantly affects poverty alleviation and economic growth [18–21]. In the context of developing countries and South Africa, financial inclusion is divided into formal and informal financial inclusion. Formal financial inclusion is extremely modernized, while the informal financial inclusion which includes stokvels, mashonisas’ etc. is socially rooted and strong [22]. Through formal and informal financial inclusion, communities that are at risk can achieve social and economic development [23]. Financial inclusion is a significant driver of sustainable development [24]. For instance, it is positioned prominently as an enabler of other development goals in the 2030 SDGs, where it is featured as a target in eight of the seventeen goals. These include SDG one, which aims to end poverty [25]; SDG two, which aims to end hunger, achieve food security, and promote sustainable agriculture [26]; SDG three, which focuses on improving health and well-being [27]; SDG four, which focuses on ensuring that all individuals have access to quality education and lifelong learning opportunities [28]; SDG five, which aims to achieve gender equality and women’s economic empowerment [29]; SDG eight, which aims to promote economic growth and jobs [30]; SDG nine, which supports industry, innovation, and infrastructure [31]; and SDG ten, which aims to reduce inequality [32]. Financial exclusion poses a danger to population health since people without access to financial services frequently lack the resources to deal with medical crises like illness, making financial inclusion essential to improving population health outcomes. Therefore, financial inclusion is expected to play a significant part in ensuring fair access to healthcare [33]. In addition, financially included individuals can participate in community-based health financing programs and save for medical costs [34]. Financial inclusion also lowers out-of-pocket costs and aids in funding healthcare, improving health outcomes and quality of life [33]. Recently, very few research has conducted empirical assessments of the impact of financial inclusion on health outcomes and mostly these studies are concentrated in Africa [33, 35–39]; in other developing countries [4, 40, 41]; and in Asia [42–44]. Even with the contribution of prior studies, some potentially important research areas still require investigation. For instance, there is limited literature that comparatively analyzes formal and informal financial inclusion’s effect on health by distinguishing formal from informal savings and credit across countries and geographical regions. Therefore, it is important for scholars and policymakers to understand whether the impact of financial inclusion on health varies between those who are formally and informally included. Against this background, this study extends and contributes to the literature on financial inclusion and health in sub-Saharan Africa by distinguishing formal from informal savings and credit using South African panel data between 2008 and 2017. Considering the above, this study seeks to achieve the following objectives: To examine the impact of formal financial inclusion (savings and credit) on health outcomes in South Africa. To examine the impact of formal financial inclusion (savings and credit) on health outcomes in South Africa. To examine the impact of informal financial inclusion (savings and credit) on health outcomes in South Africa. To examine the impact of informal financial inclusion (savings and credit) on health outcomes in South Africa. To provide actionable policy recommendations for integrating informal financial practices with formal financial systems to enhance public health. To provide actionable policy recommendations for integrating informal financial practices with formal financial systems to enhance public health. The rest of the paper is structured as follows. The next section discusses literature, followed by the discussion of research methodology and data, then presentation and discussion of the results, and the last section provides conclusion and policy implications. Literature review Theoretical framework The social determinants of health framework and the capability approach are the two interrelated theoretical stances that form the foundation of this study. Combining these perspectives provides a strong foundation for understanding how financial inclusion, especially through access to formal and informal financial services, might improve health outcomes in South Africa. The social and environmental settings in which individuals are born, grow, live, work, and age and make lifestyle choices comprise the social determinants of health framework [45]. According to [46] social determinants of health concept contend that social and structural factors including work, education, financial stability, and access to basic services influence health outcomes. In addition, the social determinants of health framework mandate that financial inclusion be measured not just in terms of access but also in terms of its resilience and usefulness [47]. Therefore, self-reported health outcomes depend on formal financial inclusion (savings, credit, and microinsurance uptake); informal financial inclusion (stokvels savings and mashonisa credit); and control variables such as education levels, household income, employment status, and geographic access to healthcare. Formal financial inclusion is expected to have a positive coefficient for health outcomes [33]. While informal financial inclusion is expected to have mixed impacts [48]. Amartya Sen created the capability approach in the 1980s to change welfare economics [49]. Most importantly, the capability approach goes away from methods that place more emphasis on how quickly a goal can be accomplished than on the objective itself. The capability approach emphasizes a person’s real flexibility to perform things that they may value doing or becoming, rather than utilities. It suggests that financial inclusion increases people’s flexibility to choose basic needs including high-quality healthcare, education, clean water, and sanitary facilities that enhance people’s health [50]. It is a normative paradigm that measures personal well-being by emphasizing human flourishing as opposed to financial gain [51, 52]. According to this viewpoint, financial inclusion improves people’s capacities by empowering them to make well-informed decisions, manage their resources, and make plans. Like social determinants of health framework, the capability approach approves self-reported health status as a measure for health outcomes rather than just life expectancy. In addition, the approach redefines financial inclusion measurement away from passive access such as having an account to dynamic capability [53]. Therefore, it captures formal financial inclusion by active usage of regulated savings, health insurance, and credit, while informal financial inclusion is captured by participation in burial societies, stokvels, or community lending. Formal financial inclusion is expected to have a positive coefficient [43]. However, informal financial inclusion is expected to have mixed impacts [48]. The social determinants of health framework and capability approach offer a convincing, multi-dimensional lens for examining how formal and informal financial inclusion impact health outcomes. Jointly, they move the focus from merely having a bank account to investigating whether financial access truly increases a person’s freedom and resources to live a healthy life. Empirical literature Empirical studies on the effect of financial inclusion on health utilize data on the proximity of financial services, such as the number of automated teller machines and commercial bank branches per 100 000 people to measure financial inclusion. For instance [36], used random and fixed effects as well as instrumental variable estimation methods to examine the effect of financial development on child health outcomes and life expectancy in 46 Sub-Saharan Africa (SSA) countries over the period 1995 to 2014. The results of both the random and fixed effects as well as the two-stage least squares estimation approaches reveal that financial development leads to better health outcomes. Similarly [38], confirms that financial inclusion measured by the number of commercial bank branches per 100 000 people leads to better health outcomes. Their study employed random effects, fixed effects, and instrumental variable fixed effects regressions to examine the effect of financial inclusion on access to basic drinking water and sanitation in Africa using a sample of 33 countries for the period 2004 to 2018. They reported that financial inclusion enhances access to basic drinking water and sanitation services. Similar results were obtained from a different estimation strategy employed in the context of African countries where population health is relatively poor. For instance [37], employed dynamic panel system generalized method of moments regression to investigate the effect of financial inclusion on population health measured by life expectancy and death rate in 33 African countries from 2004 to 2018. They reported that financial inclusion measured by the number of automated teller machines and number of commercial bank branches enhances population health. In Asian countries [43], employed two-stage least squares and generalized method of moments regressions to investigate the effect of financial inclusion on population health using panel data from 2007 to 2019. Their study confirms that financial inclusion positively impacts life expectancy and has a negative impact on the death rate in Asia. Similarly [39], employed panel regressions to examine the long-run and short-run impact of financial inclusion on health indicators such as infant health and average longevity in 29 sub-Saharan Africa nations over the period 2007 and 2018. The study reports an inverse short-run and long-run effect of financial inclusion on newborn deaths, but positive on average life expectancy. Literature that measures financial inclusion by aggregating multiple variables through principal component analysis or the global findex database into a singular index also report a positive effect of financial inclusion on health. For instance [40], measured financial inclusion by hybrid financial inclusion index and employed a pooled ordinary least squares regression model to examine the effect of financial inclusion on life expectancy and infant mortality rates on 61 developing and transitional economies in the period 2011–2017. The study reports that financial inclusion exerts a direct positive effect on health outcomes. Similarly [4], employed two-stage least square to estimate the effect of financial inclusion on life expectancy and infant mortality in 121 countries between 2004 and 2020. They also found that financial inclusion measured by applying the principal component analysis approach to four key financial inclusion variables such as automated teller machines, outstanding deposits with commercial banks, outstanding loans from commercial banks, and branches of commercial banks improves population health outcomes in countries with lower income inequality and higher gross domestic product per capita, mobile and internet penetration. In addition [44], employed Lewbel-instrumental variable estimations to examine the impact of financial inclusion on health in Singapore, Thailand, Indonesia, Philippines, and Myanmar. Their results confirm that financial inclusion positively affects health outcome, proxied by the personal health index. Research concentrating on the active use of financial systems, such as ownership of bank accounts, outstanding credit/loans, digital transaction volumes, or possession of health/micro-insurance confirms that financial inclusion enhances health. A study by [33] examines the link between financial inclusion and out-of-pocket health expenditure in Ghana using data from the seventh round of the Ghana Living Standards Survey. The study reports that financial inclusion is associated with an increase in households’ out-of-pocket health expenditure. Similar results were reported by [41] who employed fixed-effect model, two-stage least squares, and system generalized method of moments to investigate how financial inclusion can impact healthcare access in developing countries using panel data for the period 2004–2022. The study also found that there is a positive relationship between financial inclusion and health. In addition, similar results were reported in Asia by [42] who employed ordered probit model to examine the impact of credit constraints from formal financial institutions on Chinese rural residents’ health in 2014. Their results confirm that credit constraints from formal financial institutions significantly and negatively affect the self-rated health of Chinese rural residents. Furthermore [35], investigated how financial inclusion affected Nigerian heads of household’s mental health. By using a georeferenced financial services data from the Insight2Impact (i2i) GIS interface in conjunction with data from the 2015/2016 Nigerian General Household Survey, their study reports that financial inclusion significantly improves mental wellbeing. In terms of literature, there is limited evidence that comparatively analyzes formal and informal financial inclusion’s effect on health by distinguishing formal from informal savings and credit across countries and geographical regions. In addition, there is scant literature on the effect of financial inclusion on health outcomes that uses self-reported health status to measure health outcomes. Therefore, it is important for scholars and policymakers to understand whether the impact of financial inclusion on self-reported health varies between those who are formally and informally included. Research methodology and data Research data description The National Income Dynamics Study, South Africa’s first panel study, provided the data used in this research. By gathering information on socioeconomic variables including health and financial inclusion, this data source aims to monitor and comprehend the pernicious impacts of South Africa’s persistent poverty. The National Income Dynamics Study began with wave one in 2008 and has since gathered information from 7300 households and 28,000 people in five waves: wave two in 2010–11; wave three in 2012; wave four in 2014–15; and wave five in 2017. Household, adult, and child questionnaires are used in each wave to gather data on households, children, and adults. Each cycle of data collection involves fieldworkers administering these surveys to the same person. Since the National Income Dynamics Study included the study’s variables of interest such as health outcomes and its most predominant determinants such as financial inclusion, age, household income, employment status, level of education, gender, and geographic location as suggested by the literature, the study found it to be an ideal source. Different datasets are generated from data collection by the National Income Dynamics Study. The National Income Dynamics Study website enabled access and download of these datasets. Since this study concentrates on adults, the study only took samples from the adult households, individuals, and household derived datasets. The same individual can be identified across waves in these datasets using a unique identifier known as a personal identifier (pid). To determine factors and determinants, the study reviewed the adult and household questionnaires, and data from each wave. Using the pid, the study combined the datasets in each wave before adding all the waves to a single dataset. After excluding participants who did not participate in all five waves, the total number of observations was reduced. To address selection bias caused by attrition, the study estimated random effects ordered probit model which assumes unobserved individual effects are uncorrelated with the independent variables. Key variables measurement Dependent variable The most popular format for health measure is “in general, would you say your health is”, with the response options “excellent”, “very good”, “good”, “fair”, “poor”. In this study responses from this question were reversed coded so a higher score reflects better health. The reliability of reverse coding health scores is high because it maintains consistent directionality [54]. For instance, in the context of this study as financial inclusion increases, health outcomes should also show a clear, positive increase. Early research that examined the connection between health and socio-demographic factors used self-reported health outcomes to measure overall health [55, 56]. Health in this study is measured using self-reported health outcomes: 1 “poor”, 2 “fair”, 3 “good”, 4 “very good”, 5 “excellent”. Explanatory variable Following [57], financial inclusion is measured by access/usage of formal and informal financial services. Therefore, a dummy value of “1” was respectively allotted if the individual has access to formal savings and credit, and “0” otherwise. Similarly, savings of stokvel and loan of moneylender (mashonisa) are allotted a dummy value of “1” if accessed and “0” otherwise. Control covariates The rest of the variables are as per [58]: level of education = 1 “no schooling”, 2 “primary (grade R to 7/standard 5)”, 3 “secondary incomplete (grade 8/standard 6/form 1 to diploma with less than grade 12/Standard 10)”, 4 “matric (standard 10/grade 12/form 5)”, 5 “post-secondary (certificate with grade 12/standard10 to higher degree (masters, doctorate)”; employment status = 1 if “employed”, and 0 “otherwise”; household monthly take-home income = 1 “zero/nothing”, 2 “less than R3000”, 3 “between R3001 and R6000”, 4 “between R6001 and R12000”, 5 “between R12001 and R24000”, 6 “more than R24000”; age = 1 “18 years and above”, 0 “otherwise”; gender = 1 “male”, 2 “female”; and geographic area = 1 “rural areas such as farms/communal/traditional lands”, 2 “urban areas”. Econometric estimation strategies and empirical model specification The ordered probit model considers the discrete, ordinal nature of self-reported health. The ordered probit model has been widely employed in a variety of contexts in health research and shown to be more appropriate than ordinary least squares estimation [59, 60]. The random effects ordered probit specification is possibly the best approach for panel data on self-reported health outcomes as it takes into consideration the existence of an additional normally distributed cross-section error term. The cross-section error (see below) captures factors such as stable, unmeasured personal characteristics that are known to influence self-reported health. There is no time series variable which would adequately reflect stable, unmeasured personal characteristics across individuals. Therefore, the cross-section error term acts as a correction for neglecting these factors in the model. This study provides empirical evidence on the impact of formal and financial inclusion on health outcomes. A random effects ordered probit estimation strategy is adopted to investigate the impact of formal and informal financial inclusion on health outcomes. Based on theoretical and empirical evidence discussed in the preceding section, financial inclusion influences health outcomes. Let be the latent, unobservable health outcomes of individual i at time t. The reduced-form equation to be estimated is expressed as: Where: : dummy variable = 1 if the individual has access to formal savings and credit, and 0 otherwise; : dummy variable = 1 if the individual has access to stokvels and informal credit (mashonisa), and 0 otherwise; = vector of time-varying and time-invariant control variables such as level of education, employment status, household income, geographical location, age, and gender; = the individual-specific random effect; and = the idiosyncratic error term. This method assumes that individual unobserved heterogeneity is uncorrelated with the independent variables . Since is an unobserved latent variable, this study only observes the categorical ordinal outcomes mapped through estimated cut-points : The k represent thresholds to be estimated (along with coefficients) using maximum likelihood estimation, subject to the constraint that . In Table 2, the estimates of the four threshold parameters (k) are significant at 1% and 5% levels, indicating that respondents meaningfully distinguish between the five standard health outcomes (e.g., excellent, very good, good, fair, poor). In line with closely related literature by [61]. Table 2Random effects ordered probit modelVariablesCoefficientsStandard errorsFormal financial inclusion0.0497***0.00916Informal financial inclusion-0.0522***0.0135Level of education0.154***0.00496Employment status0.0665***0.00905Household income0.0259***0.00396Age-0.0244***0.000298Gender-0.133***0.00911Geographical area-0.0818***0.00905Cut points /cut1-2.827***0.0292 /cut2-2.005***0.0280 /cut3-0.954***0.0273 /cut4-0.0676**0.0271 /sigma2_u0.1494723***0.0055468 0.13003558LR test vs. oprobit model: chibar2 (01) = 1209.76 Prob > = chibar2 = 0.0000* Significant at 10% level** Significant at 5% level*** Significant at 1% level Random effects ordered probit model * Significant at 10% level ** Significant at 5% level *** Significant at 1% level In the random effects model, is assumed to have the composite nature: is the unobserved, time-invariant, individual specific heterogeneity [62]. It is assumed that is unrelated to any explanatory variable, producing the random effects model. is assumed to follow independently normal distributions with a zero mean and constant variance . is a white noise error term. It is assumed that is normally distributed with zero mean and unit variance . Since is present in the composite error in each time interval, the are serially correlated across time. In fact, under the random effects assumptions: is to be estimated. is the proportion of the panel-level variance component to the total residual variance. Likelihood ratio (LR) statistics, with one degree of freedom, are used to compare the random effects ordered probit model with the simple ordered probit model. Specifically, the test will be for the null hypothesis that . When the panel-level variance component is important, and the random effects ordered probit represents a significant improvement on the simple ordered probit model. To control for the time dimension of the panel data, a series of year dummies were included in the model but found not to materially affect the inferences (hence not included in the results as presented). The first equation is also estimated as a simple ordered probit model, ignoring the random effects. In this case, it is assumed that the error term is normally distributed with zero mean and unit variance. This allows for comparisons of the statistical inferences about the explanatory variables across the two versions. It can then be inferred whether including random effects matters for drawing other conclusions about the explanatory factors that play important roles in modelling self-reported health outcomes. Results and discussion Table 1 presents summary statistics for the dataset. Data on self-reported health outcomes shows people with mostly positive health outcomes: 4.3% report poor, 10.2% fair, 27.1% good, 28.7% very good, and 29.7% excellent health. In general population health surveys, where most healthy people report positive results, this skewed distribution is quite common [63]. A generally healthy population is suggested by the high density in the “excellent” and “very good” categories. Data on formal financial inclusion shows that on average 38.5% of the surveyed sample use formal financial institutions for their personal financial needs. In addition, 7.6% for informal financial inclusion shows a low dependence on informal financial mechanisms. This low percentage implies that most of the population surveyed in the sample depends on formal financial institutions. Level of education data shows that most of the individuals 41.3% within the sample have dropped out before completing their high school. Household income data reveals that most of the people, 38.3% of the surveyed dataset, fall into the low-income category of earning less than R3,000 per month. The average age of the individuals in the surveyed dataset is 39 years, dominated by females 57.9% residing in urban areas (54.4%). Table 1Descriptive statistics of the data sampleVariableObsMeanStd. Dev.MinMaxSelf-reported health Poor88,2130.0430.20301 Fair88,2130.1020.30201 Good88,2130.2710.44401 Very Good88,2130.2870.45201 Excellent88,2130.2970.45701 Formal financial inclusion106,8510.3850.48701 Informal financial inclusion106,8510.0760.26601Level of education no schooling96,9120.1060.30801 primary96,9120.1980.39801 secondary incomplete96,9120.4130.49201 Matric96,9120.1650.37101 Post-secondary96,9120.1180.32301 Employment status106,8510.3290.4701Household income Zero92,17000.01601 Less than 300092,1700.3830.48601 Between 3001–600092,1700.2810.4501 Between 6001–12,00092,1700.1910.39301 Between 12,001–24,00092,1700.0930.29101 More than 24,00092,1700.0510.22101 Age106,58439.69316.91618111Gender Male106,8310.4210.49401 Female106,8310.5790.49401Geographical area Rural101,9920.4560.49801 Urban101,9920.5440.49801 Descriptive statistics of the data sample The parameter estimates, estimated cut points, estimated proportion of the panel-level variance component to the total residual variance (), and estimated panel-level variance component labelled sigma2_u results of the random effects ordered probit model are presented in Table 2. The study finds that formal financial inclusion dummy has a significant positive impact on health implying that individuals with access to formal financial services have a higher probability of reporting better health outcomes. These findings align with the social determinants of health viewpoint that access to formal financial services provides the necessary resources to escape poverty-related health traps [64]. Therefore, by just increasing income is not sufficient if individuals lack savings and credit infrastructure to protect their finances during medical emergencies. In addition, these results confirm the capability approach argument that access to formal financial services gives individuals the power to choose and access healthcare exactly when they need it [50]. Therefore, an individual’s financial capability transforms passive economic resources into actualized health outcomes. Most recent research by [4, 39, 44] also confirms these results. Informal financial inclusion dummy has a statistically significant negative impact on health. This implies that higher use or reliance on informal finance shifts individuals toward the lower, less desirable outcomes of health. These results sharply challenge empirical literature and fundamental assumptions of social determinants of health framework and capability approach by implying that access to exploitative loans or high-interest informal groups can be a cause of material deprivation and chronic stress, making inclusion a health risk factor rather than a protective social determinants of health. Level of education has a statistically significant positive impact on health suggesting that higher educational attainment systematically shifts individuals into better self-reported health outcomes, after controlling for time-invariant unobserved characteristics and individual-specific random effects. This outcome is consistent with the results of [65], suggesting higher education for higher health outcomes. Employment status variable has a statistically significant positive impact on health suggesting that being employed increases the probability of an individual reporting a higher outcome of health. This outcome aligns with the results of [66], implying that being employed is associated with higher healthcare access. Household income variable has a statistically significant positive impact on health suggesting that as an individual’s household income increases, they are significantly more likely to report being in a higher, better health outcome. This outcome is consistent with the results of [67], arguing that income is positively associated with self-rated health. Age variable has a statistically significant negative impact on health suggesting that as individuals get older, their underlying latent health outcomes decreases. Gender variable has a statistically significant negative impact on health implying that women are inherently more likely to report lower health outcomes than men. Age and gender results support the outcomes of [1] for the case of the Kingdom of Saudi Arabia. Geographic area variable has a statistically significant negative impact on health implying that individuals residing in urban geographic areas have a lower underlying latent health status compared to those in the rural areas. This outcome is consistent with the results of [68], because rapid and unplanned urbanization lead to negative social and environmental health impacts, which hit the poorest and most vulnerable the hardest. The value, which represent the proportion of the panel-level variance component to the total residual variance is not equal to zero implying that the random effects ordered probit represents a significant improvement on the simple ordered probit model. LR statistics confirm this result and show that there is enough variability between health outcomes to favour random effects ordered probit regression over a standard ordered probit regression. Table 3 presents results of the simple ordered probit model for health outcomes without considering random effects. All main outcomes are found to be consistent in terms of direction and significance with the results of Table 2. Table 3Simple ordered probit modelVariablesCoefficientsStandard errorsFormal financial inclusion0.0582***0.00851Informal financial inclusion-0.0484***0.0127Level of education0.140***0.00422Employment status0.0826***0.00824Household income0.0266***0.00358Age-0.0228***0.000251Gender-0.123***0.00763Geographical area-0.0843***0.00768Cut points /cut1-2.637***0.0246 /cut2-1.876***0.0237 /cut3-0.900***0.0232 /cut4-0.0709***0.0231* Significant at 10% level** Significant at 5% level*** Significant at 1% level Simple ordered probit model * Significant at 10% level ** Significant at 5% level *** Significant at 1% level Conclusion This study contributes to knowledge by employing random effects ordered probit model to examine the impact of formal and informal financial inclusion on self-rated health outcomes using South Africa’s National Income Dynamics Study panel dataset for the period 2008 to 2017. Formal financial inclusion is found to be statistically significant and positive, while informal financial inclusion is negative and statistically significant. Level of education, employment status, and household income are found to significantly positively influence self-reported health outcomes, while age, female gender, and urban geographical area are found to significantly negatively influence self-rated health outcomes. Employing the random effects ordered probit model sheds light on the importance of considering the cross-section error, which captures the individual-specific heterogeneity for self-reported health model estimated using data from National Income Dynamics Study. Besides the knowledge contributions, the results established in this study are key for improving health outcomes especially in Sub-Saharan Africa. The results imply that formal financial inclusion could serve as an instrument for enhancing population health outcomes. This study, therefore, calls for policies and strategies that could enhance formal financial inclusion to improve health outcomes in Sub-Saharan Africa. Enhancing formal financial inclusion requires policymakers, financial regulators and financial institutions to address structural barriers that limit the effectiveness and efficiency of the financial system. Specific policies and strategies that could promote competition in the financial system, financial literacy, financial innovation, cost-effective delivery of financial services, and minimising financial access barriers such as collateral security are fundamental for enhancing formal financial inclusion. In addition, the results imply that informal financial inclusion could lower population health outcomes. This study further calls for policies and strategies that could integrate informal savings groups (stokvels) and informal money lenders (mashonisa) into the formal digital economy. This can provide secure, low-interest access to credit and formal health-related goods. Limitations of the study This study makes a significant contribution to knowledge and policies to improve health outcomes. However, this study is not free from limitations, which present some avenues for future studies. This study examines the impact of formal and informal financial inclusion on health outcomes such as subjective self-reported health while not considering objective child mortality and maternal mortality, and health-seeking behaviour such as immunisation and out-of-pocket expenditure. Therefore, future studies can extend this study by investigating the impact of formal and financial inclusion on objective health outcomes, immunisation, and out-of-pocket expenditure in developing countries. In addition, the National Income Dynamics Study data runs to 2017, and the South African financial inclusion landscape has changed considerably since then. Despite the surge in digital adoption, informal systems like stokvels remain highly vibrant in the country. Today’s findings would likely show a significant hybrid model of financial management, where consumers use formal finance together with informal finance. Therefore, future studies can extend this study by investigating the impact of formal and financial inclusion on health outcomes using recent dataset. Furthermore, both formal and informal financial inclusion variables are binary dummies, and they do not capture access to insurance due to data unavailability. Therefore, future studies can extend this study by investigating the impact of formal and financial inclusion on health by including access to insurance to financial inclusion variables, use continuous measures of financial inclusion such as depth or usage intensity. Lastly, the study acknowledges that the relationship between financial inclusion and health can be reciprocal. For instance, while access to savings and credit can improve health by smoothing consumption and reducing stress, individuals with good health are generally more economically active, face fewer medical expenses, and are more likely to be financially included. Therefore, future studies can extend this study by investigating the impact of formal and financial inclusion on health using econometric strategies that can address endogeneity resulting from possible reverse causality. Acknowledgements Not applicable. Authors' contributions All three authors Lwando Mjacu, Vikela Liso Sithole, and Palesa Makhetha-Kosi wrote the main manuscript text, prepared tables, did a literature review, ran the data, and wrote the analysis for the result. All authors reviewed the manuscript, wrote recommendations, ensured the statistics model was appropriate, reviewed the literature review, and reviewed the analysis for the model. Funding Not applicable. Data availability The data that support the findings of this study is available from the SALDRU https://www.nids.uct.ac.za/. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The data that support the findings of this study is available from the SALDRU https://www.nids.uct.ac.za/.