Technology for Social Good - funding African education through Big Data insights
- Summary:
- NGO Opportunity International is using Big Data from work in Uganda to undertake risk profiling and encourage mainstream commercial banks to support community schools.
In fact, research published by the World Bank reveals that for every additional year a child spends in primary school, their eventual wages rise by 10-20% per year. An extra year in secondary school is likely to see earnings jump even higher by 15-25% per annum.
Put another way, a woman without an education in Uganda might expect to earn the nation’s average annual income of $490 per year. But if she completes secondary education, her earnings potential leaps to between $2,205 and $6,693, which is a life-changing sum both for herself and her family.
However, education also has the added advantage of not only boosting income, but leading to a drop in the number of child marriages, decreasing birth rates and improvements in life expectancy. Nathan Byrd, head of global programmes at Opportunity International (OI), a non-governmental organisation (NGO) that is seeking to end world poverty, explains:
Education is a great equaliser as it helps people to get on in life…They have more self-determined futures as they gain the confidence and decision-making skills to shape their own futures. There’s also a huge social component – their kids are healthier because they know how to seek medical care and learn about health in the home context….They also ensure their own family is educated, which helps to break the spiral of poverty.
But according to the Unesco Institute for Statistics, 263 billion children around the world are currently not in school. This is despite the fact that one of the United Nations’ Millennial Goals was to ensure primary education was universal by 2015. But such ambition ultimately crumbled under the weight of inconsistent foreign aid, hidden costs, poor systems planning and a lack of available schools and school places, especially in rural areas.
As a result, despite huge initial interest and impressive enrolment levels, drop-out rates tend to be high - Unesco says that in Uganda alone, they stand at about 68%.
The problem is that, although governments in sub-Saharan Africa, for example, dedicate on average 18% of their public funds to education, a huge 43% of the region’s population are under the age of 15. This high percentage means that governments are all too often unable to keep up with demand for quality, affordable education.
Low-income borrowers
Therefore, the burden tends to falls on individual families, many of whom live below the poverty line. The expectation is that they cover an average of 30% of primary education expenses, which range from school fees and teachers’ salaries to uniforms, books and food, but many simply cannot afford it.
So OI’s mission is to provide microfinance to parents in the shape of school fee loans and to schools in the form of infrastructure and educational improvement loans. With the backing of philanthropists, donors and impact investors, the NGO works with 12 microfinance institutions in 10 countries in order to make the money available where it is needed on the ground.
In fact, at the end of March, OI’s EduFinance (OEF) programme was dealing with almost 45,000 loans worth just over $21.4 million that are benefitting nearly 450,000 children around the world – and repayment rates on these loans currently stand at an impressive 99%. Byrd explains why:
On the school fee loan side, it’s parents’ number one desire to send their children to school. You don’t have to convince them – they’ll do whatever they can to give their kids a good future. So they’ll service the loan before anything else, and we also structure our products to say ‘as long as you pay it back, you can get another loan and so can access credit at the right time for you’. Many community schools, on the other hand, struggle to exist, but if they default it hurts their reputation and makes them less sustainable so they don’t do it.
In Uganda, a key market for OI, the average loan size for families is $240, while for educational institutions it is more like $10,000. But mainstream banks have traditionally steered clear of such investment due to the small margins and perceived high risks.
Other key issues here is that low-income borrowers do not tend to have a profile in local credit bureaux and their loans are often not significant enough to warrant a credit check anyway as the cost of pulling their data together is simply too high to justify doing it. Byrd says:
A key challenge we face is a misunderstanding of the financial viability of the low-cost market by the commercial sector as well as a lack of knowledge about potential borrowers.
But research undertaken by OI, which currently provides about $3 million in loans to both families and educational institutions in Uganda each year, suggests that annual demand in the country actually amounts to more like $534 million. Local microfinance organisations currently have the capacity to support $57 million in loans, but this figure is still a far cry from what is needed.
Open platform credit model
Therefore, to try and rectify the situation, the NGO decided that it needed to tap into the resources of mainstream Ugandan banks - and that data was the way to do it. As a result, with the help of a grant from the charitable Bill & Melinda Gates Foundation and the data modelling skills of South African fintech company MyBucks, it developed an “open-platform credit model for education lending”.
This open platform, which is currently in pilot, is essentially a big database that includes up to 10 years of Ugandan loan information. Using a data mining algorithm, the system is able to understand the characteristics of borrowers versus defaulters and predict how low income families or schools aka potential loan applicants will behave when given money, thereby replacing the need for credit checks.
The system, which was developed using Shiny and is based on RESTful web services, can either be accessed online or run offline mode at the bank’s site on a free-of-charge basis courtesy of the Gates Foundation grant. Using a proxy model, it provides the financial institution with an immediate score based on a candidate’s risk profile. The goal is to make it available over time as an open source platform.
But OI has also created a comprehensive manual and training programme for banks that sign up to its scheme, which is due to launch by the end of the year. Three banks have already agreed to get involved, with room for two more. As Byrd says:
There is interest because we’re offering banks a unique growth opportunity. We can say, ‘Here’s a sector that you’re not reaching and we can help you do that’.
But Byrd also acknowledges that OI is “playing the long game here”, not least because getting the banks onside requires a lot of help and support. He explains:
Once they’re comfortable with the idea, we have to do a lot of work to learn how the bank operates. We then run a technical assistance project to create a functional algorithm for them and to integrate our models into their broader business strategy. Sometimes, it’ll be a quick process, but at other times it can take up to a year or more. It’s then about providing short- and long-term training for employees on systemised marketing, post-lending follow-up and client outreach so that they don’t have to reinvent the wheel.
My take
OI chose Uganda as the place to launch its open platform credit model as it is a post-conflict state with a large population split between urban and rural. The idea is that if it can get the scheme to work there, it is likely to be able to get it to work elsewhere in sub-Saharan Africa too. And this is an important point for an NGO that is keen to start a “movement. As Byrd points out: ‘This is just the beginning for us….”