Ethics of microfinance institutions: a study applying the radical affinity index Open Access

Microfinance has emerged as a key tool in the fight against poverty, promoting financial inclusion in marginalised communities. The initial focus on its economic benefits led, for a significant period, to the assumption that microfinance initiatives were effective and positive.

However, this perception has changed over time, partly due to evidence of problems which showed that not all programmes were beneficial (Dichter and Harper, 2007). Situations of over-indebtedness and bad practices, especially in the case of Andhra Pradesh in 2010, where accusations of misconduct by microfinance institutions (MFIs) were accompanied by a series of borrower suicides (Adbi, 2023; CGAP, 2010), have led to a perception of crisis and uncertainty in the sector (Hudon and Sandberg, 2013). As a result, there has been an increasing emphasis on the detailed evaluation of social impact or social performance, as well as the approaches and practices of these activities.

In recent years, certain aspects specific to the microfinance sector have been analysed, such as attention to financial literacy, prioritising disadvantaged groups, properly explaining interest rates and prices and designing adapted products. Alongside these specific topics, the debates share concerns with general issues of ethics in finance (customer protection, asset allocation, transparency, participation, etc.).

However, analysis of their ethical implications has not received the same rigorous attention, and the development of literature in this area is still in its early stages (Babalola et al., 2022). This paper addresses this gap by critically examining the ethical dimensions of microfinance operations and their impacts on beneficiary populations. Specifically, the aim of this study is to determine whether MFIs are ethically responsible, identifying the characteristics which lead to better performance in this regard.

Once the necessary adaptation to the microfinance sector has been carried out, the methodology of the radical affinity index (RAI), initially used for traditional and ethical banking (San-Jose et al., 2011), is applied to a significant sample of MFIs. The MFIs included in the latest available list from the MIXMarket database, with both financial and social performance information, are assessed using an adapted version of the RAI to determine the ethical level of their practices. From this information, three relevant aspects that can influence their ethical levels are analysed: the type of entity, the geographic area and the size.

This paper contributes to the current debates on microfinance by allowing, with this ethical perspective, the establishment of guidelines towards aligning the purpose of MFIs with the humanistic goals for which they were created.

The paper is organised as follows. In Section 2, the ethical aspects of microfinance are explored, presenting a theoretical review of the issues it addresses and the applicable ethical perspectives. Section 3 summarises the index used to determine the ethical perspective of microfinance and its adaptation to the specific circumstances of these financial entities. In Section 4, the hypotheses to be tested are presented in relation to the main characteristics of microfinance, addressing the type of entity, its location and its size. This is followed in Section 5 by a detailed account of the sample data and the methodology used. Section 6 presents the results along with the discussion. The paper concludes with a summary of the conclusions, the limitations of the study and potential lines for future research.

Microfinance, in its current form, emerged in the 1970s in Bangladesh and Latin America (Hudon, 2011). Its expansion to different regions occurred especially from the 1990s onward and was driven by international organisations (United Nations, World Bank and others), development cooperation agencies and philanthropic organisations. The positive image of these practices as a means of fighting poverty was reinforced by declaring the year 2005 as the International Year of Microcredit and the awarding of the Nobel Peace Prize to Muhammad Yunus in 2006 for his contributions in this field.

However, this positive image has become more nuanced over time as practical problems and unexpected, undesirable results have emerged. Today, microfinance is a mature sector with highly diverse practices, making the evaluation of its social impact both appropriate and necessary.

By their very nature of unsecured lending to the poor and the small amounts granted, microfinance interest rates are usually higher than those in traditional banking, although they are much lower than those of local moneylenders or loan sharks. In practice, this can be described as a “poverty penalty”, which paradoxically means that the poor pay more for access to financial services (Gutiérrez-Nieto et al., 2017).

In this context, there have been some notorious cases of abuse. For instance, in 2007, the owners of the Mexican MFI Compartamos made excessive profits by selling part of their stake. In this case, the MFI’s good results were linked to interest rates of over 100%, much higher than the usual rates in the sector (Lewis, 2008). There have also been various situations of promoting over-indebtedness or using unethical practices for credit recovery (Hulme and Arun, 2011). The most extreme case resulted in shocking suicides in India (Adbi, 2023; CGAP, 2010).

From another perspective, the feminisation of debt is questioned (Mayoux, 2002). Since women are in a situation of greater vulnerability and poverty, microfinance programmes target them as clients, but this is not always positive. In practice, the credit obtained by women benefits the household as a whole and, in many cases, the decisions on its use are made by men. This feminisation of debt, where the woman must take responsibility for the repayment, is considered by some authors (Nilakantan et al., 2021) as an “ethical violation” facilitated by many MFIs.

From a teleological standpoint, various authors question the effectiveness of microfinance (especially microcredits) as a solution to poverty (Bateman, 2010; Roodman and Morduch, 2014) and, increasingly, there are demands for evidence of their positive impacts. In reality, it is clear that credit cannot be solely considered a tool for financial empowerment, but it also involves risk-taking and future indebtedness by people with limited resilience to failure. In this sense, encouraging risk-taking among people in precarious or exclusionary situations is dangerous for them and creates an obvious moral hazard (Guérin et al., 2014).

Much of the criticism stems from the predominance of microcredit as virtually the only microfinance instrument, despite the dangers of indebtedness and the fact that it is not always what recipients need to improve their living conditions. In fact, there is an increasing focus on microfinance and financial inclusion, with an approach that goes beyond credit to offer all types of financial services (savings, payment methods, insurance, etc.) tailored to low-income populations (Collins et al., 2009; Demirgüç-Kunt et al., 2022). This perspective aligns with that of the bottom/base of the pyramid approach, which advocates for the possibility of alleviating poverty by serving poor sectors in a profitable manner, and has generated significant debate since the beginning of the century (Dembek et al., 2020).

It is also argued, setting certain projects aside, that the push for microfinance is strategically embedded in the global political economy, facilitating the liberalisation of the financial sector on a global level (Weber, 2004).

However, when looking more closely at the practices of the entities, problems of mission drift can often occur as initiatives grow and consolidate with commercialisation processes. In general, mission drift is understood as shifting to a new segment of clients who are in a better financial situation or to existing clients who have been successful and want to increase their loan amounts (Mersland and Strøm, 2010). This represents an MFI’s deviation from its original intention of serving the most disadvantaged, thus contravening a fundamental universal ethic in MFIs (Serrano-Cinca and Gutiérrez-Nieto, 2014; Mia and Lee, 2017). This deviation, often measured by the size of the loans as a proxy for the target clients, does not occur in all cases. It can also be viewed positively, such as in cases where an MFI begins to provide services to less poor individuals to achieve commercial scale (Kleynjans and Hudon, 2016). Beyond mission drift, the potential conflict between social and financial objectives is an area of research that continues to generate debate in the sector (Awaworyi Churchill, 2020; Gutiérrez-Goiria and Goitisolo, 2011).

Although from a generalist ethical perspective the range of possibilities may be broader, in the studies related to MFI, three fundamental perspectives of ethical analysis have been identified (Chakrabarty and Erin Bass, 2015):

1. deontological;

2. consequentialist; and

3. the so-called virtue ethics, which is also known as humanist (Alzola, 2012).

These perspectives are summarised in Figure 1.

Based on these three perspectives, Table 1 shows the emphasis, application and normative purpose that each implies.

In most studies, the ethical considerations regarding MFIs are approached by assessing their outcomes from a utilitarian perspective, evaluating the impact on the improvement of people’s quality of life or the socio-economic environments in which they operate. The dual social and financial objective of MFIs (Hudon, 2011) necessitates a dual analysis, examining both the efficiency in resource use and the effectiveness in enhancing the lives of the beneficiaries, while avoiding the exclusive use of financial criteria as indicators of success.

The efficiency of MFIs concerning stakeholders, particularly users, must be maximised. An optimal Kaldor–Hicks solution, where some benefits compensate for the losses of others, is insufficient; instead, a Pareto optimal solution, where no one is made worse off, is desirable. The social objective requires not only instrumental outcomes but also final outcomes, aiding in the personal and collective development of communities. While this line of research is relevant, it is rooted in a consequentialist or utilitarian approach. There are other possible frameworks, notably deontological and humanist, which are the focus of this study.

The commitments of microfinance cannot be a mere subsidiary of traditional finance, as “the promise of microfinance was founded on innovation: new management structures, new contracts, and new attitudes” (Morduch, 1999: 1572). From a humanist ethical perspective, any evaluation must focus on the well-being of the beneficiaries as the ultimate goal (Tisdell and Ahmad, 2018). This paper proposes that consequentialist and humanist approaches can align, promoting a holistic ethical perspective (Sathye et al., 2014). However, the desirable coexistence of “doing good” and “doing business” does not always occur (Karnani, 2011), as evidenced by numerous documented conflicts (Chiu, 2014).

The magnitude of these conflicts and the risk of mission drift (Aubert et al., 2009; Copestake, 2007; Mersland and Strøm, 2010; Xu et al., 2016; Serrano-Cinca and Gutiérrez-Nieto, 2014) necessitate a deeper ethical reflection on MFIs, not only from a results perspective but also from a coherence perspective. To what extent are the strategies, processes, practices, actions and attitudes of MFIs consistent with the social change they aim to promote?

Three aspects for reflection, addressed in part by Hudon and Sandberg (2013) are:

1. General ethical justification for microfinance: Why should it exist?

2. Appropriate behaviour of MFIs: How should they behave?

3. Characteristics of the microfinance environment and provider: What is the appropriate structure and strategy?

The first issue is approached from a macroeconomic or philosophical perspective, analysing the ethical function of MFIs in the socioeconomic context from various viewpoints. The second issue focuses on analysing cases related to the actions of the entities or their agents. Both lines of inquiry are essential for advancing knowledge.

However, this paper focuses on the third issue: studying characteristics of entities that may align with more ethical behaviours, improving the well-being of beneficiaries. Although it has implications for the second issue, it is centred on the internal characteristics of MFIs.

From a humanist ethical standpoint, appropriate performance is first associated with prioritising purpose (Hudon, 2010; Pogge, 2005) and then with developing an ethical framework of interrelationship among stakeholders (Labie, 2007). An MFI is ethical if there is alignment between its purpose and its actions, similar to the affinity approach (Cowton, 2002), used in ethical banking (San-Jose et al., 2011). Outcomes and purpose are necessary but not sufficient to define the ethical character of an MFI. A utilitarian analysis can instrumentalise people for other objectives, and the risk of mission drift, as previously mentioned, is significant. Aspects such as transparency, effective asset placement, guarantees and the degree of stakeholder participation, especially users, define the ethical character of MFIs.

San-Jose et al. (2011) classify banks based on their adherence to ethical principles and propose an index, the so-called RAI. The RAI can offer a novel approach to examining MFIs by evaluating their ethical commitments and social orientation.

This index is particularly insightful for distinguishing between ethical and traditional banks. The RAI can also be reformulated to highlight the ethical view of other types of financial entities, as demonstrated with Islamic banks (San-Jose and Cuesta, 2019). In applying the RAI to Islamic banks, the unique characteristics of these institutions were incorporated into the ethical measurement index. Specifically, the placement of assets evaluates whether the banks’ investments align with social and environmental priorities, giving preference to those that contribute positively to society. Guarantees assess the inclusivity and fairness of loan security mechanisms, emphasising the use of alternative guarantees to promote financial inclusion. Participation measures the involvement of various stakeholders, including Shari’ah supervisory boards, in the banks’ decision-making processes, ensuring broad representation of interests and adherence to Islamic ethical standards. Thus, the RAI is tailored to reflect the unique ethical principles of Islamic finance, emphasising both Shari’ah compliance and broader social responsibilities.

The purpose of microfinance is to provide financial services to low-income individuals or those without access to typical banking services. MFIs aim to promote financial inclusion by offering credit, savings, insurance and other financial products to underserved populations, thereby fostering economic development and reducing poverty. The four components of the RAI – transparency, placement of assets, guarantees and participation – are particularly useful for highlighting the ethical dimensions of MFIs, ensuring they meet their social objectives effectively and ethically. The following is a brief explanation of the four components adapted for microfinance.

Transparency is crucial for stakeholders to monitor and assess the behaviour and impact of MFI’s activities. While transparency is important in both microfinance and traditional banking, microfinance also presents unique characteristics in this respect. In microfinance entities, transparency should encompass accountability not only in financial aspects but also in social and mission-related issues, which are central to the sector. In this context, the MIXMarket platform, supported by leading microfinance agents and institutions, has been a significant source of information on the global performance of MFIs. Its inclusion as one of the World Bank’s databases (World Bank, 2025) gives an idea of its relevance. A key aspect of transparency in microfinance involves the terms of products, especially loans. Small loan amounts lead to high costs and consequently high effective rates. Additionally, fixed costs (such as opening fees) have a significant relative impact on these small amounts. Financial illiteracy and short, small payment terms further complicate clear information for clients and other stakeholders. This concern led to the creation of the MicroFinance Transparency initiative in 2007, aimed at providing clear information on the prices charged to the poor by the industry. After eight years of addressing financial problems and lack of commitment and participation, the initiative ceased in 2015 (Waterfield, 2015). Alongside these voluntary initiatives, regulations have been developed in various countries, requiring MFIs to publicise specific aspects of their activities.

Microfinance emerges with the purpose of facilitating financial inclusion and combating poverty by serving traditionally excluded sectors. In this context, the average loan size (in monetary units or as a proportion of per capita income) is a usual indicator (Awaworyi Churchill, 2020; Gutiérrez-Goiria and Goitisolo, 2011; Gutiérrez-Nieto et al., 2017; Memon et al., 2022). A smaller average loan size reflects a focus on individuals with fewer resources, allowing an MFI to be identified as truly oriented towards the poorest individuals (who are largely women and rural inhabitants). This indicator is also used in studies analysing potential mission drift, where some MFIs may gradually shift towards serving wealthier clients as they mature (Copestake, 2007; Mersland and Strøm, 2010; Mia and Lee, 2017; Serrano-Cinca and Gutiérrez-Nieto, 2014; Xu et al., 2016).

In microfinance, guarantees play a crucial role in mitigating risks associated with potential loan defaults, serving as a source of repayment should such defaults occur (Stiglitz and Weiss, 1981; Armendáriz de Aghion and Morduch, 2005). However, in MFIs, guarantees are primarily used to persuade rather than reduce risks, as MFIs aim to reach individuals lacking resources and collateral. Innovative systems, such as group guarantees where members act as mutual guarantors, enable disadvantaged individuals to access credit (Müller et al., 2014). Despite the association of MFIs with reduced collateral requirements (Islam, 2016), doubts remain regarding their collateral-free operation. Research suggests conventional guarantees are unsuitable for MFIs, with alternatives like social sanctions and credit denial proving effective (Bond and Rai, 2002). Furthermore, Müller et al. (2014), in their study on credit risk in agricultural microfinance, find minimal effectiveness of conventional guarantees. In summary, while guarantees are crucial in MFIs, their role differs from traditional finance, focusing on innovative and inclusive approaches to support underserved populations.

Incorporating robust participation mechanisms in MFIs is essential for enhancing inclusivity, accountability and sustainability. Stakeholder participation means that donors, employees, customers or creditors take part in an organisation’s governance – especially by being members of the board – to help monitor managers, share information and improve performance (Mori and Mersland, 2011). Currently, MFIs often do not involve clients, particularly women, in decision-making processes, treating them as passive recipients rather than active participants. This lack of participation limits their empowerment and integration. By involving borrowers and community members in governance, MFIs can ensure that financial services align with the actual needs of the community, fostering trust and stronger client relationships. Such inclusivity not only enhances transparency and accountability but also drives innovation, as diverse perspectives contribute to more effective financial solutions. Research by Memon et al. (2022) empirically examines the impact of female participation on the financial sustainability and outreach of MFIs, revealing that while female participation improves social performance, it may decrease financial sustainability. This underscores the need for MFIs to balance social goals with financial health. Adhering to ethical standards that promote stakeholder involvement can enhance the reputation of MFIs and attract international support.

The concern about ethics in MFIs extends beyond their inherent ethical standards to identifying the conditions or characteristics that enable better outcomes and promote good ethical practices in the sector. To address the gap in the literature, this study explores the factors that influence the ethical perspectives of MFIs. Consequently, the study examines three hypotheses, based on previous literature, which address the legal situation, geographical context and size of MFIs.

Among MFIs, we find both NGOs and non-banking financial institutions (NBFI), which are characteristic of this sector, as well as traditional entities such as banks, credit cooperatives or rural banks. The importance of the type of entity in various aspects of microfinance has been frequently studied. For example, Gutiérrez-Nieto et al. (2017) find significant differences regarding key operational aspects, affecting the interest rates offered in each case, which paradoxically may be higher when serving lower-income populations, as seen with many NGOs. Awaworyi Churchill (2020) shows that for-profit MFIs outperform not-for-profit MFIs in terms of financial sustainability, but not-for-profit MFIs are better in terms of depth of outreach. Similarly, Gutiérrez-Goiria and Goitisolo (2011) find that NGOs excel in their social orientation. However, Adbi (2023) finds that, contrary to the general assumption of better financial performance of for-profit MFIs, non-profits can be more sustainable, at least in the context of a scandal such as the 2010 Indian microfinance crisis. Meanwhile, Serrano-Cinca and Gutiérrez-Nieto (2014) show that NGOs are better positioned to serve the so-called long tail, composed of clients needing smaller amounts, which are traditionally less profitable, aligning directly with the social perspective of microfinance. In summary, the type of entity can be a key aspect for studying ethical behaviour, with this behaviour expected to be more apparent in entities with a strong social mission. This leads to H1:

Microfinance, in its current approach, emerged almost simultaneously in Asia and Latin America and continues to have the greatest presence in these regions, although it has expanded to areas such as North Africa and the Middle East, Eastern Europe and Sub-Saharan Africa. The divergence in practices between geographic areas is significant and stems from issues such as the proportion of women involved or the average loan sizes used (Memon et al., 2022; Tisdell and Ahmad, 2018). Gutiérrez-Goiria and Goitisolo (2011) find the lowest profitability in Africa and notable differences in terms of profitability and orientation between South Asia (with a greater focus on women and working with small amounts) and Latin America, suggesting the coexistence of different microfinance models. Geographic differences are often introduced in studies as dummy variables (Gutiérrez-Nieto et al., 2017) and can provide some comparative insights for the future direction of the sector. In this regard, the H2 states:

Microfinance operates on a relatively small scale in terms of loan amounts and client numbers, but some MFIs grow to be quite large, with millions of clients. Maintaining commitment and focus as these initiatives expand is a common challenge for entities with a strong social mission, and microfinance is no exception. Research on mission drift often considers scale as a significant factor, measured through indicators such as loan portfolio size (Copestake, 2007; Xu et al., 2016) or total assets (Serrano-Cinca and Gutiérrez-Nieto, 2014; Mersland and Strøm, 2010). Additionally, Memon et al. (2022) include size as a control variable in their analysis of women’s participation and MFI sustainability. Size is also pertinent in discussions of potential trade-offs between objectives (Awaworyi Churchill, 2020; Gutiérrez-Goiria and Goitisolo, 2011). This study examines whether the size of an MFI is related to its ethical behaviour, specifically whether larger entities can sustain high ethical standards. Following this approach, H3 states:

The study uses the extensive standardised global MIXMarket database, recently incorporated among the World Bank databases (World Bank, 2025).

The sample consists of entities from this database that provide both financial performance data (Financial Performance data set) and social performance data (Social Performance data set) for the most recent years available at the time of the study (fiscal years 2019 and 2018, respectively). This is a set of 79 entities with a proven track record and capacity, which also show a commitment to accountability in different aspects (382 entities provided financial data, and 171 social data, but only 79 of these did so in both cases). To enhance the information, a review of the entities’ websites was conducted from October to December 2023, which resulted in the exclusion of entities that had ceased operations or undergone significant changes in orientation. Ultimately, the analysis focuses on a sample of 62 entities.

To evaluate the ethical behaviour of entities, the methodology developed by San-Jose et al. (2011) is adapted. This approach analyses entities across four key areas, scoring them according to the criteria detailed below, with a maximum of four points in each area:

Table 2 shows the scores for transparency in MFIs. Starting from a base of 1, the maximum score would be 4 if the three additional aspects were adequately disclosed.

The aim here is to assess whether assets are allocated to areas of specific interest. Firstly, the proportion of assets that the MFI allocates to its loan portfolio is taken into account. Additionally, a multiplier ranging from 0 to 4 is applied to this value, depending on the loan amounts granted. Thus, the average loan balance per borrower in relation to gross domestic product (GDP) per capita is used as an indicator. This widely used value indicates the type of client the MFI is targeting. Lower values indicate better asset allocation, as they target populations with fewer resources and a higher probability of being excluded. The formula used is as follows:

Placement of assets value = Multiplier*(Loan Portfolio/Assets).

Multiplier_i = 4* (Maximum average loan balance per borrower in relation to GDPpc − average loan balance per borrower in relation to GDPpc_i)/(Maximum average loan balance per borrower in relation to GDPpc − Minimum average loan balance per borrower in relation to GDPpc).

The multiplier takes a value of 4 if the average loan balance per borrower in relation to GDP per capita is the minimum in the sample (0.95%) and a value of 0 if the average loan balance per borrower in relation to GDP per capita is the maximum in the sample (394.18%).

Microfinance has relied on innovative methodologies since its inception; these replacing the usual guarantees of the traditional banking system. In reality, however, very different practices can be found which facilitate (or hinder) access to finance, depending on the case. Table 3 shows the values assigned to the different types of guarantees.

MFIs should foster greater interaction among shareholders. To achieve this, it is essential to develop participation systems for approving operational criteria and asset allocation. Table 4 shows the values given to the different forms of participation.

Once the data for each entity were tabulated and verified by at least two researchers of the team, the aggregate data and results were analysed by category, including legal status and geographic area. In addition to providing overall descriptive statistics, t-tests for equality of means were performed, without the need to assume equal variances (homoscedasticity was not required). Finally, an ordinary least squares regression model was tested, with the calculated RAI value serving as the dependent variable, and the values of total assets, average loan balance per borrower in relation to GDPpc, legal status and region as explanatory variables.

Figure 2 shows the value of RAI and its components in the entities analysed. As can be seen, transparency (3.03 on average, with a maximum of 4) is the most highly valued aspect according to the study criteria, followed by placement of assets and guarantees. Participation is the component with the lowest value.

By type of MFI, Table 5 shows some interesting differences. Firstly, the highest RAI value is obtained on average by NGOs, which obtain comparatively high scores in 3 of the 4 components, while the lowest is obtained by rural banks, which together with banks obtain scores below average in all sections.

To observe in detail the differences between entities from an ethical point of view, Table 6 shows the differences in means between the cases. Leaving aside the Credit Union and Other categories, which only have one entity, t-tests of equality of means are carried out, without assuming equal variances.

As shown in Table 6, NGOs are the highest-rated entities overall. Both NGOs and NBFIs achieve a significantly higher RAI compared to banks and rural banks, with rural banks receiving a significantly lower RAI than the other categories. With some nuances, this situation is also observed when disaggregating the components of the RAI.

With regard to transparency, the rural banks studied are rated significantly lower than the other categories. For reference, the highest-rated cases (4 points, 19 entities among the 62) not only include economic and social information from the database but also have a website that clearly specifies their products and conditions, informs about the use of resources and provides information explaining the overall functioning of the entity in its different areas.

With regard to placement of assets, based on the capacity to grant small loans, NGOs are rated significantly higher than other entities. Conversely, rural banks perform significantly worse than the others in this area, consistent with other aspects.

For guarantees, no significant differences are observed. In practice, the type of entity does not seem to be strongly related to the guarantees required, based on the available information. It is noteworthy that most entities (47 out of 62) include the possibility of providing guarantees beyond traditional ones, such as other types of assets, personal experience or track record or group or community guarantees.

Participation is highlighted as being the lowest-rated aspect and is an issue that should be reconsidered by MFIs if they truly want to advance towards functioning with high ethical standards that promote the involvement of stakeholders. As a reference, no true system of stakeholder participation in governance is found in any of the cases studied. The most common form of participation (17 cases) involves some system of volunteering or involvement in activities. This aspect differs from other ethical banking activities, where participation in matters such as loan orientation is relatively common. Overall, this aspect shows the greatest significance in the differences, with NGOs promoting the most participation, followed by NBFIs, banks, and finally, rural banks.

As observed in Table 7, entities from Latin America, Eastern Europe and Central Asia and East Asia and the Pacific are the highest rated on average, with only minor differences among them. These are followed by South Asia and the Middle East and North Africa. Sub-Saharan Africa is in last place, with a significant difference in the RAI and its components of transparency and participation ( Appendix).

A multiple regression analysis was conducted to examine the joint relationship between several predictor variables and the dependent variable, RAI (average of the four measures). As in the previous analysis criterion, the two categories with only one institution in each case (Credit Union and Other) were excluded. Some key aspects of the institution were included as predictors: its assets, its target (average loan balance per borrower in relation to GDP per capita) and its legal status. Sub-Saharan Africa was also included as a dummy variable, given that the previous analysis indicated a possible effect associated with this geographical issue.

Dependent variable:RAI (average of the four components).

Predictors: Constant term, Total assets (US$), Average Loan Balance per borrower in relation to GDPpc, a dummy variable for NGOs (Dum_NGO), a dummy variable for NBFIs (Dum_NBFI), a dummy variable for rural banks (Dum_Rural bank) and a dummy variable for Sub-Saharan Africa (Dum_Africa).

As Table 8 shows, the R-value of 0.633 indicates a high correlation between the predictor variables and the dependent variable. The R-squared value of 0.400 suggests that approximately 40% of the variability in the dependent variable can be explained by the model. The adjusted R-squared value of 0.331 adjusts for the number of predictors in the model and provides an accurate estimate of the true explanation given. The standard error of the estimate is 0.4433, indicating the average distance that the observed values fall from the regression line. Finally, the Durbin–Watson statistic of 1.694 suggests that there is no significant autocorrelation in the residuals of the model.

The constant term is statistically significant (B = 2.511, p < 0.001), indicating a significant baseline level of RAI. Among the predictors, the average loan balance per borrower in relation to GDP per capita shows a significant negative effect on RAI (B = −0.004, p < 0.001), suggesting that higher average loan balances per borrower are associated with lower RAI scores.

The dummy variable for rural banks (Dum_Rural Bank) is also significant (B = −0.552, p = 0.022), indicating that rural banks tend to have lower RAI scores compared to the reference category (Banks). Additionally, the dummy variable for Sub-Saharan Africa (Dum_Africa) is significant (B = −0.405, p = 0.026), implying that organisations based there tend to have lower RAI scores compared to their non-African counterparts.

Other predictors, such as Total assets, Dum_NGO and Dum_NBFI, did not show statistically significant effects on RAI.

Collinearity diagnostics indicate that multicollinearity is not a concern in this model, with all tolerance values well above 0.1 and variance inflation factor (VIF) values below 10.

The analysis of RAI and its components provides valuable insights for policymakers and stakeholders in the microfinance sector, suggesting aspects of the entities’ characteristics that may require targeted interventions to improve organisational performance and RAI scores, particularly for rural banks and Sub-Saharan African entities.

Out of those reviewed, transparency was the highest-scoring aspect. Of the 62 entities evaluated, 45 were in the top two levels of the scale, which was to be expected given their financial and social accountability to MIXMarket. In fact, many entities provided very comprehensive information about their programmes, products and conditions, the lack of which has previously been a criticism of the sector (Waterfield, 2015).

Allocation of assets was the next best-rated aspect, though it showed the greatest variability, including extreme cases. A focus on excluded sectors with small-scale operations is a hallmark of the sector and represents a key aspect for studying potential mission drift (Copestake, 2007; Mersland and Strøm, 2010; Mia and Lee, 2017; Serrano-Cinca and Gutiérrez-Nieto, 2014; Xu et al., 2016).

The availability of guarantees demonstrates the sector’s willingness to move beyond traditional collateral, with various examples of alternatives, including group guarantee loans, in line with literature that suggests the need for different types of guarantees (Bond and Rai, 2002; Müller et al., 2014; Poornima et al., 2025).

Finally, participation was clearly the lowest-rated factor, indicating a significant shortfall in this type of programme. Participation is crucial for innovation in microfinance (Morduch, 1999) and can contribute to its social objectives (Memon et al., 2022). Furthermore, participatory mechanisms can prevent mission drift, keeping MFIs focused on their social mission rather than prioritising profitability. While female participation is crucial and offers significant rethinking, other stakeholders such as governments, non-female clients, depositors and the broader community should also be involved to ensure all perspectives are considered in MFI decisions.

With this general overview in mind, an assessment is now made of the hypotheses previously raised:

H1 is partially confirmed. The analysis shows a higher average RAI value in NGOs and NBFIs compared to banks and rural banks. The differences in means are significant, with NGOs being the most prominent entities in having a greater ethical commitment. The only credit union in the sample obtains values similar to those of NGOs and NBFIs.

Of the RAI components, the aspect of transparency is significantly worse in rural banks, while NGOs excel in participation and placement of assets. This latter aspect is linked to their focus on the poorest segments, a key aspect in microfinance, thus confirming studies along this line (Gutiérrez-Nieto et al., 2017; Serrano-Cinca and Gutiérrez-Nieto, 2014). No significant differences are observed in guarantees.

The multivariate analysis, which includes factors such as the size of the entity and its geographical area, confirms the poorer performance of rural banks. In the case of NGOs and NBFIs, although the values are positive, they are not significant. The lower average loan balance per borrower in relation to GDP per capita could partially explain the better ethical performance of these entities.

The hypothesis is confirmed. The analysis shows geographical differences in the RAI, with Sub-Saharan Africa showing a notably poorer performance. Latin America shows high values in transparency (3.5 out of 4), with entities providing detailed information about their operations.

The analysis indicates that, from an ethical standpoint, the observed differences do not replicate those seen from other perspectives (Gutiérrez-Goiria and Goitisolo, 2011; Memon et al., 2022; Tisdell and Ahmad, 2018). The situation in Africa suggests the need to delve deeper into regulatory aspects or factors that could be improved.

The hypothesis is rejected. There is no significant relationship between the size of the entities and their ethical behaviour. This suggests that additional factors beyond size influence the ethics of MFIs. Large entities may implement mechanisms to maintain their ethical commitment despite their size, or there may be variations that neutralise the effect of size.

In summary, the results obtained reveal a significant relationship between the main variables related to a holistic humanistic ethic, transparency, placement of assets, participation and guarantees, analysed through the RAI and some of the dimensions studied. Both the legal status and the geographical context influence the ethical responsibility of microfinance entities, but not the size, measured in volume of assets. NGOs and NBFIs show a greater ethical commitment, especially in participation and placement of assets, while rural banks score poorly on transparency. Participation remains a critical area for improvement to align the practices of MFIs with their social and ethical objectives. There also seems to be a need for greater imagination in relation to the guarantees used to reduce risks.

The data show that a greater ethical commitment on the part of MFIs may lead them to opt for more social legal forms, such as NGOs and NBFIs, or at least demonstrate a certain reluctance to move to commercial legal formulas. The poorer ethical performance of rural banks in terms of transparency suggests an analysis is necessary to evaluate to what extent this is motivated by their organisational specificity or by the different realities they address. In any case, improvements should be considered in all four dimensions in the case of rural banks, given that their values are low both in absolute terms and in comparison with other institutions. Clearly, it is impossible to change the geographical area, nor can a specific geographical environment be discounted. Nevertheless, it is possible to identify the best practices in entities that maintain a stronger ethical commitment and aim to imitate them.

Regardless of the impact that an increase in ethical responsibility may have on results from an instrumental perspective, it is clear that microcredit entities, both because of their public positioning and their dealings with disadvantaged individuals, are morally obliged to develop high ethical standards, something that they do not always achieve. The data obtained can help nudge them in the direction of social and human commitment, without them necessarily having to forgo efficiency in terms of results, either with regard to their own sustainability or to the social impact achieved.

The study confirms that NGOs and NBFIs exhibit higher ethical standards compared to banks and rural banks. This is evident in their higher average scores for RAI and its components, particularly in participation. Placement of assets, while well-rated, shows considerable variability, reflecting the sector’s focus on serving excluded sectors with small loans – a key indicator for studying mission drift. Guarantees are another positive aspect, with the sector moving beyond traditional collateral to include group loans, aligning with literature advocating alternative guarantees. However, participation remains the weakest area, indicating a significant gap in involving clients, especially women, in decision-making processes, crucial for innovation and achieving social goals.

The study has limitations, including a sample limited to entities reporting to MIXMarket, potentially biasing results towards more transparent organisations. The cross-sectional design does not account for changes over time and some predictors, like total assets, do not significantly influence RAI scores. Future research should include longitudinal studies to assess the evolution of ethical practices and expand the sample to non-reporting entities for a more comprehensive view. Qualitative studies on low participation rates and strategies for client involvement, especially in Africa, are essential. Structural equation modelling methodology will be useful in future specialised studies, given its robustness for non-normal distributions, small samples and other statistical problems. Additionally, exploring the impact of digital financial services on transparency and participation could provide valuable insights.

The authors would like to thank the editor and the anonymous reviewers for their constructive comments and helpful suggestions, which significantly improved the manuscript. This research was supported by the University of the Basque Country (UPV/EHU) under projects GIU 21/011, GIU22/003 and US25/12.