1. Introduction
In the context of financial inclusion, the phrase ‘gender gap’ is commonly used to denote the unconditional mean difference in outcomes between men and women. This difference is often equated with ‘gender-based discrimination’, a term frequently used in an expansive sense- sometimes to refer to intended and unintended forms of discrimination[1]in the way financial products are designed and/or delivered, and at other times to denote barriers arising from gender norms. However, insofar as these differences are unconditional, i.e., they don’t control for other observable characteristics that could be contributing to those differences, they do not capture gender-based differences, whether arising from discrimination, norms, or other factors. Attributing such unconditional gaps solely to ‘gender’ is therefore misleading.
What then constitutes a gender gap? To answer this question, we examine both unconditional and conditional gender differences in financial inclusion across three dimensions- access, usage, and financial well-being. We use the India data from the 2025 Global Findex dataset to restrict this analysis to the Indian context. We run two types of regression, which we refer to as the ‘unconditional’ and the ‘conditional’ specification.[2] In an unconditional specification, we estimate unconditional gender differences, capturing raw differences in financial inclusion outcomes between women and men. In the conditional specification, we estimate conditional gender differences while controlling for other factors, based on the available data. In other words, the conditional specification compares men and women who are similar in all observable respects, isolating gender as the only distinguishing factor.
In the following sections, we summarise the results from this analysis and discuss our interpretation of these results. The details pertaining to the hypothesis, regression model, analysis results, and a list of further readings can be found in our Appendix document.
2. Findings
2a. Access
In the unconditional specification, we find evidence of gender differences across almost all access variables except account ownership. While there is no statistically significant difference between women and men in basic account ownership, women are significantly less likely than men to own other financial instruments. The unconditional gender difference is 18 percentage points (pp) for mobile money and digitally enabled accounts, 15 pp for debit cards, 3 pp for credit cards, 5 pp for savings at a bank and owning an insurance account, and finally between 3 pp and 6 pp for borrowing via formal institutions.
In the conditional specification, controlling for age, education, income quintile, region, employment status, and mobile access, we find that gender differences either disappear or reduce in magnitude and significance in most instances. This points to the importance of socio-economic characteristics as key channels that influence these gender differences. Women are 8 pp less likely to have a mobile money account and a digitally enabled account, and 4 pp less likely to have a debit card. Moreover, statistically significant differences in access to insurance, formal borrowing, and formal savings disappear. Interestingly, conditioning for other observables makes women 5 pp more likely than men to own a bank account. Reviewing the control variables, we note that higher education, higher income quintiles, employment, and mobile phone access are strongly and positively associated with most access indicators.
2b. Usage
In the unconditional specification, we find evidence of gender differences in the usage of financial accounts, as women are significantly less likely than men to report usage across most digital and account-based financial activities. Specifically, women are 11 pp less likely to make or receive a digital payment, 10 pp less likely to use a bank account for savings, 12 to 14 pp less likely to use a mobile for making a payment or checking account balance, and 30 pp less likely to use a credit card. Similar to our access results, the gender difference decreases in magnitude when other control variables are accounted for. For instance, the gender difference in using a bank account for savings and using a mobile for making digital financial services transactions almost reduces by half. Moreover, the conditional gender difference in making or receiving digital payments is statistically insignificant.
The control variables exhibit strong associations with usage outcomes. Education shows large positive gradients across nearly all indicators, particularly for digital payments, online purchases, and account management activities. Income gradients are similarly pronounced, especially at higher quintiles, with higher-income individuals substantially more likely to engage in digital usage. Employment is positively associated with most usage indicators, particularly digital payments and transaction activities. Mobile access is strongly and positively associated with nearly all digital usage outcomes, underscoring its central role in enabling engagement.
2c. Well-being
Questions on financial well-being, even though administered to a single individual in the Findex survey, directly or indirectly implicate the ‘household’. The necessity of making this distinction arises because the questions are framed to understand ‘how long a household could cover expenses if the main source of income was lost’, ‘formal savings for old age’, and ‘difficulty in arranging emergency funds’. It would be fair to say that financial decisions and actions, such as managing emergencies and planning for retirement, are typically made at the household level and have implications for the family as a whole. Therefore, individual responses to questions on financial well-being should not be read as gender differences in perceptions of ‘individual-level’ financial well-being, but rather as gender differences in perceptions of ‘household-level’ financial well-being. This distinction is particularly salient in India, where approximately 95 per cent of households comprise multiple members, reinforcing the analytical importance of the ‘household’ as the relevant unit for studying financial well-being.
In reviewing gender differences in financial well-being, we focus on three variables: saving formally for old age, raising lumpsums for emergencies, and household resilience to income loss.
In the unconditional specification, women are 11 pp less likely than men to report saving formally for old age. This difference reduces by 3 pp in the conditional specification. Among the control variables, income shows the strongest and most consistent association with formal old-age saving. In the unconditional specification, women are significantly more likely than men to report difficulty in mobilizing emergency funds within 30 days. Women are about 15 percentage points more likely to report that accessing emergency funds would be very difficult, and correspondingly, 5 pp less likely to report no difficulty. In the conditional specification, we find evidence of gender differences in raising emergency funds, although the difference is smaller when compared to the unconditional specification. Among the control variables, education and income show strong associations with emergency fund resilience, while age, region, employment status, and mobile access do not seem to influence the ability to raise emergency funds.
Finally, we find evidence of both unconditional and conditional gender differences in perceptions of household resilience to income loss. Women are 6 pp more likely to report funds lasting for only less than 2 weeks and 5 pp less likely to report funds lasting greater than 2 months, if the main source of income was lost. These estimates show little change between unconditional and conditional specifications, suggesting that observable socioeconomic characteristics account for very little of the gender-associated differences. This possibly indicates that psychometric factors could explain most of the gender differences in this dimension of financial resilience.
3. Discussion
3a. The ‘what and why’ of gender gap
Our results indicate that gender differences in the unconditional specification are driven by multiple factors besides gender itself. Subsequently, when we control for these factors, we find that gender differences either disappear in some instances or reduce in magnitude in most instances. It is only when a statistically significant gender difference remains even after controlling for other observables, that it is appropriate to characterise such a difference as ‘gender gap’.[3]Indeed, our results indicate significant gender gaps in access to digitally enabled accounts, usage of financial products, especially digital financial services (DFS), and perceptions of financial well-being. What could explain this gap? One possibility could be missing variables due to the limited scope of data available in the Global Findex dataset. While our conditional analysis accounts for an individual’s socio-economic characteristics that influence levels of financial inclusion and well-being, it does not capture several other important determinants of financial behaviour. These include household-level factors (such as household size and source of household income), socio-cultural factors (including intra-household roles and responsibilities and social capital), and psychological factors (typically captured by the Big 5 personality traits), all of which have been shown to shape financial decision-making. Moreover, the data does not capture supply-side variables such as the availability of financial infrastructure and the quality of financial services available, which are equally important determinants of financial access and usage.
However, the other possibility is that even if this data were available and included in our analysis, a significant gap would persist, potentially due to two reasons – (i) men and women who are otherwise similar receive the same product/service, but because the product/service is designed for men rather than for women, women are less likely to own and use the product/service; (ii) men and women who are otherwise similar are subject to different social and cultural norms, and it is these norms that prevent women from owning and using formal financial products/services. Below, we explore these two potential reasons for the said gender gap.
3b. A case of bad design
A lot of the gains made in women’s access to formal financial services are grounded in gender intentional design. The deployment of women agents, women’s SHG program, and the many women-based cash transfer schemes have all collectively contributed to women’s increased access to bank accounts and other basic banking services. However, for certain other kinds of products, both access and usage remain muted and are cases of bad design. The formal financial system typically offers a standardised set of products/services to its customers, which are designed to suit men’s lived realities. This leads to an inadvertent exclusion of women because women do not want the same products/services as men do. Take for instance, access to digitally enabled accounts and use of such accounts for making digital payments, where we find a significant gender gap. Existing research tells us that fear of fraud and inadequate grievance redressal mechanisms in DFS play on women’s minds a lot more than men, holding them back from using DFS. Gender intentionality, in this context, would involve designing and delivering DFS in ways that improve women’s trust and confidence. Gender intentionality, therefore, requires understanding the ways in which women are different from men and accounting for those differences in the design and delivery of products/services. Enabling women’s “equal access” to formal finance, therefore, does not mean offering the same/identical products/services but rather offering non-identical or non-same products/services to women relative to men. However, once we acknowledge and accept such a notion of “equal access”, the goal of “equal usage” does not retain any conceptual content. If this goal were still maintained, then we should expect to find (and not be surprised) that “equal access” will not lead to “equal usage”.
3c. Treating norms as given versus norms as obstacles that need overturning
If norms are indeed what is causing a gender gap, it is worth asking if the norm can be easily changed without any harmful consequences for the men and women in question. If the answer is yes, then we can see if changing the norm will in fact result in greater ownership by women, and if it does not, then norms are unlikely to be the underlying driver of the gender gap, and we are back to a case of bad design, as in 3b above. If the answer, however, is no, then it means that the product/service will have to be redesigned to ride alongside the norm rather than against it.
A growing body of research supports the view that product innovation and policy interventions should align with existing social norms rather than attempt to overturn them. For instance, Moscona et al. (2026)[4], writing about development policy interventions in general, argue that efforts to change norms often rely on fragile assumptions and may generate unintended consequences; instead, interventions that take social and cultural traits as given and design within those contexts are more likely to succeed. Evidence from the specific developmental arena of financial services also bears out the more general result. Kusimba (2018)[5] shows that the success of mobile money in Kenya can partly be attributed to the gender norm of relational work (the effort women invest in building and maintaining social relationships through financial exchanges), suggesting that existing norms can sometimes facilitate, rather than hinder, financial inclusion. In the same vein, studying rural Niger, Thomas et al. (2025)[6] identify a “culturally wise” model of agency grounded in social harmony, respect, and collective advancement. Through a series of field experiments, they show that psychological interventions aligned with this interdependent model of agency significantly improve household economic outcomes, whereas interventions based on more individualistic, Western notions of agency do not produce similar effects.
In essence, these examples highlight that norms need not be seen as barriers, and even if they are seen as barriers, it may be more effective to design alongside them rather than against them. In other words, the question of whether the norms can be changed need not be the first question to ask once the existence of norms has been acknowledged. Gender intentionality in this context would mean designing and delivering financial products and services in a manner that aligns with existing gender norms. This could mean, for instance, designing financial products and services for women based on their intra-household roles and responsibilities. If this entail money management in the form of managing household budgets, juggling debt, building relational savings, then any product/service that helps women manage these responsibilities effectively, could have a better chance of increasing their levels of engagement with formal finance and closing the gender gap. Here too, it is worth pointing out that “equal access” to formal finance through gender intentional design, such that it rides alongside gender norms, means designing non-identical product/service for women and men to match the norms that they are subject to, respectively. And so, once again, the notion of “equal usage” has little conceptual content once the non-identical nature of products/services has been acknowledged.
Footnotes:
[1] By ‘intended’ form of discrimination, we mean a deliberate effort by the financial system to exclude women, whereas by ‘unintended’ form of discrimination, we are referring to the inadvertent exclusion of women due to the nature of product design and delivery. A bank officer rejecting a woman’s loan application on the basis of her gender constitutes an intentional form of discrimination. A woman unable to access an ATM because of its distance is an unintentional form of discrimination.
[2] In the paper titled ‘Access to finance in Sub-Saharan Africa: is there a gender gap’, the authors make a similar distinction and differentiate between conditional and unconditional gender gap in access to finance.
[3] This is well understood in the economics literature. Nobel laureate, Claudia Goldin in this podcast on gender pay gap explains that when comparing men and women with similar human capital characteristics doing equal hours of work, any statistically significant gender differences in wage can be characterised as gender pay-gap.
[4] Moscona, J., Nunn, N., & Robinson, J. A. (2026). Searching for Fish in Trees? Economic Development when Context Matters (No. 2026-27).
[5] Kusimba, S. (2018). “It is easy for women to ask!”: Gender and digital finance in Kenya. Economic Anthropology, 5(2), 247-260.
[6] Thomas, C. C., Premand, P., Bossuroy, T., Sambo, S. A., Markus, H. R., & Walton, G. M. (2025). How culturally wise psychological interventions can help reduce poverty. Proceedings of the National Academy of Sciences, 122(46), e2505694122.
