The Power of Microdata From Global Findex

Measuring financial inclusion is key to identifying opportunities to overcome barriers that prevent people from accessing, using, and benefiting from financial services. The Global Findex database, launched in 2011 by the World Bank, has made it possible to measure financial inclusion worldwide in a systemic and comparable way from the perspective of the end-users of financial services. Is it possible to use its granular data to examine financial inclusion at the individual level?

This demand-side survey offers numerous opportunities for policymakers, researchers, donors, and financial inclusion practitioners to better understand how people interact with the financial system, both regulated and unregulated. As part of CGAP’s research on the impact of policy responses to financial crises on financial inclusion outcomes, we used granular data from Global Findex to measure the frequency of access to and usage of financial services, drawing from individual responses to 13 questions. The Findex microdata gathered in 2011, 2014, 2017, and 2021 covers nearly 150,000 individuals from over 140 countries. Microdata is yet to be released from the 2025 Findex. Retaining individual-level data helps inform evidence-based policymaking by revealing insights about vulnerable groups, enabling more targeted and effective interventions.

Measuring the frequency of access to and usage of financial services

Most analyses using the Global Findex dataset have relied on aggregated data with a binary nature, such as having an account at a financial institution or not. Would it be possible to build metrics that capture differences between individuals in their degree of access to and usage of financial services?

Our indicator is an attempt to do just that. To this end, we identified 13 questions that are recurring across the 4 waves of the Global Findex Data. The more questions a respondent answered affirmatively —indicating access to or usage of financial services (e.g., debit accounts, debit cards, credit cards, loans, deposits, withdrawals, mobile payments, electronic payments, etc.)— the higher the score assigned to that individual. Affirmative answers are aggregated, and then the indicator is normalized to range from 0 to 10, with higher values indicating greater depth and breadth of financial inclusion. Our score incorporates traditional aspects of financial inclusion, specifically the usage and access dimensions. However, it does not capture other relevant issues like financial health. Despite this, it offers a practical basis for integrating additional layers into a continuous indicator, rather than using numerous separate questions.

When combining many factors into one indicator, traditional statistical methods like principal component analysis can be hard to interpret, as their goal is to maximize explained variance. By contrast, our indicator is designed to be simple: a higher value means a person has more access to and uses more financial services. Each of the 13 questions is weighted equally, so no single type of access or usage counts more than another. Because the score has a fixed scale, it highlights differences between individuals, making comparisons easy.

In all four waves many people scored zero – those who consistently answered "no" to all questions. For those with a score greater than zero, the levels of financial inclusion can vary significantly. Figure 1 shows the financial inclusion score for each country over a ten-year period, using data from the Global Findex surveys conducted in 2011 and 2021. Not surprisingly, the score varies significantly across the globe and has shown improvement over time.

Regions such as North America and Europe exhibit higher values of our indicator, whereas countries in Latin America and Sub-Saharan Africa tend to have much lower levels. While this geographic disparity is well-documented, our indicator provides an effective way to summarize this information through a single continuous variable, making global comparisons across individuals more straightforward.

We present our score disaggregated by income, gender, and age – three demographic dimensions consistently included in each wave of the survey. Each figure reports the scores calculated for 2011, 2021, and the full sample period (2011, 2014, 2017 and 2021). As Figure 2 shows, there is a clear positive trend across income quintiles; higher income levels are associated with higher financial inclusion scores. Importantly, there has been significant progress over time: while in 2011 the median score for income quintiles below the 80th percentile was zero, by 2021 this was no longer the case, with all quintiles showing a median score above zero.

Data disaggregated by gender also provides interesting insights (See Figure 3). Across waves, women tend to have a median score of zero, whereas men show a higher median score, indicating greater financial inclusion. Interestingly, when comparing 2011 and 2021, gender disparities have relatively narrowed.

Figure 4 suggests that the age-related gap has also diminished over time. Scores for the working-age population (ages 15–64) have increasingly aligned with those of individuals aged 65 and older, indicating a convergence in financial inclusion across age groups.

The way forward: How can this indicator provide meaningful insights?

What truly sets our indicator apart is its ability to summarize multiple dimensions of financial access and usage while preserving the richness of granular data. Unlike aggregated measures at the regional or national level, we can identify the most excluded sub-populations whose experiences are often obscured when indicators are averaged. This opens new possibilities for analysis.

For example, rather than estimating the average gender gap in financial inclusion, we can now ask: How does this gap evolve as overall financial inclusion improves? Granularity becomes valuable when assessing the impact of policies, as it avoids the flawed assumption that all individuals respond identically to the same intervention. Using granular data empowers researchers and policymakers to understand the dynamics of the most vulnerable, offering insights that can drive more targeted and effective policy solutions.

To give a preview of what we will share in the coming months, we are employing this indicator to better understand how past policy responses to major systemic crises affected financial inclusion. In particular, distinguishing between those individuals excluded from the financial sector and those who are increasingly included, and thus, determining whether the standard policy toolkit may inadvertently exacerbate existing gaps. Stay tuned!

Note: Country-level indicators are calculated using weights. The microdata for the 2025 wave was not accessible at the time of this blog's publication.