Mapping Africa Financial Inclusion- Results Review and Next Steps

The landscape data yields the following big numbers as outlined earlier: 71 million clients, 44 million deposit accounts, 20 million loans, across 23,000 providers in 45 countries. If we begin to break this data out by products and provider types, we can see that:

• Credit unions serve the most in the aggregate, although most are small and many sectors have little or no presence from this model
• Most sectors have a diverse set of providers, covering a range of institutions offering different products - no single model dominates overall
• Mobile banking reaches many in the aggregate, but we only have information on a few products with broad reach

Chart 2 shows the top ten markets by total loans, for instance.

Looking at absolute numbers only provides part of the picture. We also should look at the context of individual countries. The IMF and CGAP databases on access to finance use the adult population to normalize results across countries. For this effort, since we are focusing specifically on financial services targeted to the poor, instead we use the percent of the population living below the poverty line as a basis. We can then start to identify priority markets for targeting financial services: where are the most poor unbanked clients?

Chart 3 shows a map of the sector based on the absolute gap between poor populations and financial services targeting the poor - each country is shaded by the difference between total clients and the population living under the poverty line. While this is by no means a precise estimate - people can have multiple accounts across institutions and products, and these providers may have many non-poor clients - it does provide a quick guide to the magnitude of the gap in some markets.

Using absolute numbers, we can see that Nigeria and the Democratic Republic of the Congo have the largest gaps between populations living in poverty and those with access to financial services - 80 million in Nigeria and 48 million in the Congo. (Conversely, Cape Verde and Kenya are the only markets in which the number of accounts exceeds the population living below the national poverty line (since an individual can have more than one account). For Kenya, this is in part due to the outreach of mobile banking, which has only recently been linked with savings or credit services.) In Nigeria, we can see this gap exists despite the distribution of hundreds of microfinance providers across the country.

The situation in the Congo is similar, with data on almost 200 credit unions and several dozen NGOs in our sample. Looking at relative figures, we can see that gaps are most prominent in Central and Southern Africa, as well as in post-conflict countries.

As noted earlier, savings leads most sectors. Chart 4 shows deposit-to-loan ratios by country (on a log scale, for markets with more than ten million in population); markets are color-coded by the percent of the population below the national poverty line. We can see some relatively high-poverty markets, such as Rwanda, Nigeria and Madagascar have below average deposit to loan ratios.

While the number of mobile banking projects remains small, their reach is broad and this sector covers more than 18 million people. 12.6 million of these are reached through Safaricom’s M-Pesa product, with another five million via Vodacom and MTN products. Beyond these three, outreach is limited, and we lack data on linkage to other financial services for the existing products.

A major addition of this data effort is to integrate information on informal and unregulated providers of financial services. One prominent group of informal providers are savings groups (often known as village savings-and-loan associations, or VSLAs). Data on this sector is now readily available via the SAVIX site and we have incorporated country-level figures into this effort. Savings groups reach around four million people in sub-Saharan Africa. While this is a small share overall relative to credit cooperatives and banks, in selected markets (as shown in chart 6), savings groups outreach exceeds or is comparable to that of credit unions and banks. Consequently, data on informal providers adds important nuance to understanding the options for the poor in these countries.

The data also sheds light on the distribution of services in particular markets. Earlier we saw deep trend data available on Benin via ALAFIA and a separate post showed trends and location for microfinance banks in Nigeria (formerly community banks). Using data on provincial distribution of service providers from FPM (Fonds de promotion de la Microfinance), we can similarly build a basic map of credit unions in the Congo DRC.

Utilizing the map displays helps us to further hone in on supply and demand gaps, as well as opportunities. We can use the maps to search for markets that have high mobile penetration or which have supportive business environments (as indicated by the IFC’s ‘Ease of Doing Business’ index) or which receive the most donor assistance. If we look at official development assistance (ODA), for instance, we can see a stronger relationship in chart 7 between savings mobilization and donor assistance than credit outreach, perhaps indicating that ODA has been directed at institution-building factors necessary to permit savings growth.

Gaps and open questions

While we believe this effort takes some important steps forward in measuring financial access for the poor, we are still left with several questions and gaps in the data and methodology.

Definitions: There are two sets of definitions that bedevil all similar efforts. The first issue is the more straightforward problem of definitions for the indicators: do all data providers mean the same thing when they report a loan or a savings account or a branch? What are the limits of using data sources that alternately count people, accounts and dollars? While these are thorny issues, they can be resolved. A small loss in precision will also have less effect on high-level sector estimates. The second, knottier, issue is the scope of such efforts: what should be included and excluded? We have taken an approach that attempts to be inclusive of informal and unregulated providers, but is exclusive of commercial providers that reach many, including the poor. Where should future efforts draw the line?

One special case for consideration is that of the consumer finance companies that are prevalent in Southern Africa. ‘Microlenders’ like Blue Financial Services (with affiliates in twelve markets) or Letshego Holdings (active in seven countries) provide small-dollar loans and often refer to themselves or are referred to as ‘microlenders’ or ‘microfinance institutions.’ However, much of their business is built on lending to salaried employees, often government workers, and using methods such as payroll deduction to facilitate repayment. This meets definitions of ‘microfinance’ based on name and dollar amounts, but did not meet the restriction of providing services that were designed for or likely to be used by the poor. Indeed, their products explicitly exclude those that work in the informal sector. For this initial effort, we opted to exclude such providers in general. However, there are strong arguments for both sides. Improving these landscape estimates should be viewed as an iterative, collaborative process that takes these views into consideration.

Timing and frequency of data: As we saw earlier, the ‘typical’ data point for this study is a little more than a year and a half old. While many sources are updated quickly and frequently, many others are not. The slow turnaround of data likely has several explanations. For some organizations or in difficult operating environments, data collection may not be a core activity. Sector surveys may be expensive and infrequent. In the end, data collection efforts will be strengthened if practitioners perceive the value of that data. For this effort, we hope that visualizing and opening up the source data supports on-use by practitioners.  

Limited depth of data: Data from most sources covers only a small number of data points, although there is always a trade-off between breadth and depth. Additional depth could be useful in two specific areas:

• Detailed location information can provide support to policy-makers and regulators, but also for business decisions made by practitioners. Whether data focuses on headquarters (for institutions with limited branch networks, such as credit unions and rural banks) or on branch locations (for institutions with more geographic breadth), improving the standardized disclosure of geographic information will allow deeper understanding of access beyond national borders. For instance, the South African Financial Sector Charter definition of ‘effective access’ specifies that individuals be within 20 kilometers of a service point. Without detailed branch-level and service-point location information, evaluation of such metrics is impossible. Detailed location information also allows integration with other geo-coded data sets, such as those on aid flows.
• Key indicators of risk and sustainability can add insight beyond the small number of metrics and monetary aggregates covered in most broad-based data collection efforts. Portfolio-at-risk ratios, yields or interest rates, and profit measures provide a more nuanced view on the differences and trade-offs across institutional models and markets.

Future steps

The ability to present this data relies on the incredible steps forward in transparency and disclosure that have taken place among networks, regulators, financial institutions, researchers and others over the past several years. We have better information than ever before about the supply of financial services available to the poor. We hope that practitioners and policy-makers use this data and recommend improvements to the coverage and methodology. We see a few key conclusions relevant for those working on data resources in sub-Saharan Africa:

• Focus on markets where data is less frequent.
• Find trusted data providers who demonstrate capacity for collecting, managing and disseminating information on their constituents and support harmonization around these efforts. The launch this year of the SAVIX site on savings groups is one positive example of this coordination.
• Fill gaps in reported indicators: The data available still has several gaps, requiring estimation and inference, leading to uncertainty. Client or account information deserves special focus in this vein - we should devote as much attention to counting people as to counting dollars, especially for small-balance accounts.

Related Articles:

• Overview
• Historical Efforts and MIX's New Approach
• Data Methodology