The Emerging Landscape of Demand-side Data in Branchless Banking

This is the second post in a series on the emerging branchless banking data architecture.

The first post in our series on emerging branchless banking data architecture highlighted the critical importance of data for policy and decision-making as well as the progress so far on gathering supply-side data.

Supply-side information helps us to understand the breadth of infrastructure for branchless banking in a given country – how many ATMs, POS devices, agents and so on. It is also the source for data on volumes of flows, pricing, and product terms. However, this is only half the puzzle and our understanding of branchless banking trends is incomplete without robust information on the demand side. Which clients are using which products for which purpose? What aspects of a service are they satisfied or dissatisfied with? And, perhaps most importantly, is the service having a positive impact on their general well-being?

For years, the microfinance industry focused almost exclusively on supply side data and assumed that more branches, more loan officers and more loans meant that poor people’s lives were improving. However, in the last few years we have learned that reality is more complex than what the supply-side growth would have us believe. We have to make sure that we integrate robust demand-side data to understand the full picture on financial inclusion, especially as the branchless banking industry evolves.

So, what’s already happening in terms of demand-side data collection and aggregation?

Many branchless banking providers conduct regular surveys to measure indicators such as the socioeconomic profiles of their clients, transaction profiles of clients and client experiences and preferences of product features and customer service. While supply side data measures what has already been accomplished (e.g., number of ATMs or accounts), demand side data is often used to help plan strategies going forward (new product development, change in marketing messages etc.). Collecting demand side data is often very expensive and private sector players who have done so usually want to keep it confidential to protect their commercial advantage in having collected it.

Mobile network operators are reluctant to share their number, type and volume of transactions, active customer base and overall profitability while commercial banks are unwilling to share total number of segmented customer base, overall profitability, loan portfolio, deposit portfolio and ATM transactions. National efforts to collect demand side data often produce detailed data sources but the data is not standardized from one country to the next, limiting the ability to aggregate or compare metrics. For these reasons, gathering demand-side data on a global level may be even more challenging. Efforts to collect global level demand-side data have begun and some exciting and detailed demand-side data is starting to emerge.

The figure from the last blog post organized data by source (demand- or supply-side) and depth/breadth of coverage. There are three main demand-side financial inclusion data sets (all three funded by the Bill and Melinda Gates Foundation) with a significant degree of branchless banking and mobile money information.

• Global Financial Inclusion Database (Global Findex), World Bank – Covering 148 economies, Global Findex provides 506 country-level indicators of financial inclusion summarized for all adults and disaggregated by key demographic characteristics – gender, age, education, income and rural or urban residence. There are a few questions related to mobile money that give an idea of the breadth of use of mobile phones for sending and receiving money.
• Gallup Payment Survey, Bill & Melinda Gates Foundation  - The Gallup survey is not nearly as broad as Global Findex (11 African countries although South Asian countries are being added) but provides more detailed information on the demand for payments services across these countries. The survey incorporates all payments transactions with distant counterparties (those in a different part of the country) for a nationally representative sample of 1,000 adults across a 30 day time period but excludes face to face retail transactions. Payments include domestic money transfers, international remittances, government and wage payments, payments for goods and other bills. The data can help inform both providers (to better help them understand where to target with which products) and policy makers (e.g., payment flows make up 11% of GDP in these 11 countries).
•  Financial Inclusion Tracking Surveys (FITS) and Tanzania Mobile Money Tracker Study, InterMedia and Bill & Melinda Gates Foundation – The FITS household panel surveys in Pakistan, Tanzania and Uganda use annual surveys to track 3,000 households in each country, and provides the industry with deep information on basic demographics, remittance activity and mobile money awareness and use.  These surveys are inspired by the M-PESA surveys conducted by William Jack of Georgetown and Tavneet Suri of MIT Sloan. The data is analyzed to understand not only trends in mobile money use but also relationships between mobile money use and household well-being. The Tanzania Tracker Study involved individual-level quarterly surveys conducted throughout 2012.  These surveys provide quicker-turnaround data, tracking market trends and providing analysis of awareness, uptake and use of mobile money by detailed demographic segmentations. A similar survey was conducted in Haiti in 2011.

The Global Findex, FITS, and mobile money tracker surveys are making all their data available online for the general public to analyze. The Mobile Money Data Center includes the first waves of FITS Tanzania and Uganda surveys and is structured to make data analysis very easy for anyone interested in exploring the data for themselves.

Perhaps the ‘holy grail’ of demand side data is the impact question. How can we understand whether branchless banking services are making a positive difference in client’s lives? A number of research institutions have used a range of qualitative and quantitative methods to test the impact of microfinance, most notably using randomized control trials, or RCTs. We have yet to see a robust set of experimental results in branchless banking although donors and researchers alike are considering how best to measure impact of this new service.

One exception is the Jack and Suri paper which used a natural experiment in the form of the M-PESA agent roll out to generate evidence that access to the system increased people’s ability to reach out to family and friends in an emergency thus significantly reducing the impact of negative shocks (such as severe illness, job loss, fire, or harvest failure) that could have knocked them back further into poverty.

Clearly, there is a lot of momentum to try and understand branchless banking clients as the industry progresses. This will be critical to ensure that products and services that are developed are truly meeting clients’ needs while providing a robust business case for providers. We’ll continue this conversation next week as we look more closely at the Financial Inclusion Tracking Surveys (FITS) in East Africa.

-------------Claudia McKay is a Financial Sector Specialist at CGAP and Jake Kendall is a Senior Program Officer at the Gates Foundation.