> Posted by Shannon Mudd, Coordinator, Microfinance and Impact Investing Initiative, Haverford College
This post is part of the Center for Financial Inclusion’s Expert Exchange: Building A Movement Toward Financial Inclusion by 2020, cultivating conversation around the goal of reaching full financial inclusion by 2020. For further questions about this series, write to Sonja E. Kelly, Fellow, Center for Financial Inclusion at Accion.
What factors determine where microfinance institutions (MFI’s) locate their branches – and where they do not? A group of students at Haverford College set out to explore this question, initially focusing their efforts on Uganda.
The project grew out of a conversation with Susy Cheston of CFI about who is excluded from financial services and whether maps might help us understand more about who these excluded are and the challenges in reaching them. As coordinator of Haverford’s new Microfinance and Impact Investing Initiative (MI3) I was seeking ways to engage students in research that bridged theory and practice. Recalling that our new Librarian had previously been involved in a mapping project, I returned to Haverford to see what resources we might bring to bear. I discovered that we had the right software, a skilled and supportive library and, best of all, some terrific students in the Microfinance Consulting Club who were willing to roll up their sleeves and get to work (properly fueled by pizza, carrots and seltzer water).
Why Uganda? Not surprisingly, such an effort is largely driven by data availability. We became aware of a directory of MFI branches in Uganda published by the Association of Microfinance Institutions of Uganda (special thanks to Scott Gaul of TheMIX.org!). The directory provided us with district level locations across Uganda’s 100 districts). Additional research helped us to further refine locations for a majority of the branches. We then looked for sub-national socio-economic data that we could turn into map layers to conduct our analysis. The product of our initial efforts has been posted here.
Determinants of MFI Locations in Uganda
What did we find? Highlights include:
- MFI’s locations are not uniformly distributed
- MFI’s locations tend to be in areas of greater population density
- Some areas of medium population densities largely comprised of poor households have concentrations of MFI branches, but others do not
- Past conflict areas may be associated with low levels of MFI branch locations
- Infrastructure can be important. Branch locations often follow roads. Areas with poor infrastructure, for example, low availability of safe water, have low concentrations of MFI branches.
Figure 1 below provides evidence for some of our insights. The map shows areas of high MFI concentration in a crescent-pattern along the banks of Lake Victoria stretching east and southwest from the capital city of Kampala. A second band of concentration lies along the Eastern border south of Lake Albert. Relatively few branches are located in the South between these two bands. In addition, there are few locations across the entire northern swath of the country.
Geographically dispersed clients raise the cost of microfinance provision, so population densities are likely to play a large role in determining branch locations. The shaded polygons indicate district population levels. Not surprisingly, the high concentration area of MFIs along the crescent of the shore of Lake Victoria, centered on Kampala, is also the area of greatest population (darkly shaded). The lighter shaded bands in the north and in the central region between Lake Albert and Lake Kyoga are areas of relatively sparse population and few microfinance locations. Also note the lack of MFIs in protected areas, for example, along the border between Lake Edward and Lake Albert. This information provides some insight into where MFIs locate their branches. But it also raises questions. For example, the southwestern areas have only moderate population density but have a relatively high number of MFI branch locations.
General population density is only part of the picture. While overall populations may be high, this does not mean that the population of poor people will be high. Figure 2 shows the population density of those in poverty, which are the target clients for MF operations. While the locations in the far southwest corner now seem justified, there are other areas of comparable density which have little coverage, in particular, the north central areas.
Figure 2 also indicates areas of tea production. Both the southwestern corner and areas running further north along the border with the Congo (DRC) show significant tea production. While tea was traditionally grown on large tea plantations, for several decades the country has supported small scale tea farmers with a few cooperative tea processing factories. This can be seen both south and west of Lake Edward. These small scale farmers and the ecosystem that supports them may be a source of good clients for microfinance.
Figure 2 provides an additional dimension that might explain the relative lack of MF locations in the north central region. There is a large area in the northern part of the country that experienced conflict from the operations of the Lord’s Resistance Army and the corresponding counter operations. While the low population levels were certainly a deterrent, repercussions of this conflict may still be present and MFIs may choose to avoid these areas due to the physical damage to the land or the emotional damage of the population that remains in the area.
While population densities provide information about potential clients, supply side issues, such as the costs of staffing branches, must also be taken into account. Infrastructure may play an important role in this regard. One proxy for infrastructure is safe water coverage. Figure 3 seems to indicate that branches are more likely to be located in areas of high safe water coverage and largely missing in areas where safe water is not available.
These maps are an initial effort both to determine whether mapping technologies can provide insight into MFI locations and to learn specifically about what factors may be affecting MFI locations in Uganda. Mapping technology does appear to be a useful tool of analysis. Its limitations, like so many types of analysis, are primarily due to limitations of data, whether of complete MFI location listings or of geographically oriented sub-national socio-economic data. We would like to urge the MFI associations and statistical services of all nations to continue their work in collecting such data and making it available for public use.
Our work at Haverford MI3 continues. Future mapping projects include a more in-depth analysis of Ugandan data (does the distribution of NGOs differ from that of for-profit MFIs?) and expansion of our coverage to Ghana. For information about MI3’s other activities, check out our website.
For more information, sign up for updates from the Financial Inclusion 2020 campaign.
Shannon Mudd is the Coordinator of the Microfinance and Impact Investing Initiative (MI3) at Haverford College. He is also a Visiting Assistant Professor in Haverford’s Department of Economics. Mudd’s research focuses on banking and finance, international finance, international economics, and development economics. Mudd holds a Ph.D. in Economics from the University of Chicago.
Have You Read?