For the past year, Cenfri and Finmark Trust have been partnering with NextBillion to share insights from their Insight2Impact (i2i) initiative, a five-year program that has sought to improve financial inclusion and related public policies through the smarter use of data. Their Openi2i Series has explored key learning areas from the program, ranging from program design and stakeholder engagement to digital skills development and innovation in Africa. As the series progressed, it evolved to include an additional (and unanticipated) focus: the impact of COVID-19 on financial and digital inclusion efforts.

Now, as the Openi2i Series concludes, i2i lead Hennie Bester takes a look back at the program and the learnings it has generated. In the interview below, he explores the “quantum leap” that digitalisation and data represent for emerging economies in Africa and beyond, and the opportunities and risks that presents for both providers and policymakers. 

Karen Kühlcke: Looking back to the origins of insight2impact and our initial objective regarding driving data-driven decision-making, what has changed in the last five years in terms of the appreciation for, and use of, data?

Hennie Bester: It’s almost unreal to think that five or six years ago, when insight2impact was launched, digitalisation of the African economy was something spoken about by futurists and the most forward-thinking governments. The goal of data-driven decision-making seemed visionary, but it has proven to be timely. Ironically, the slight extension of the programme took us into the pandemic and so, somewhat serendipitously, we’ve been able to witness accelerated digitalisation. 

KK: Digitalisation efforts have been significantly accelerated in response to COVID-19, but has the emphasis on digitalisation altered attitudes regarding the collection and use of data?

HB: So much of digitalisation is about data: It is an essential raw material without which efficiencies can’t be achieved. What’s interesting is that when we started insight2impact, we thought about data applications in the financial space in terms of measuring financial inclusion, but by the end of our project in Rwanda, we had moved from remotely measuring how many people have financial services, to leveraging data to drive policies that determine the level of financial inclusion. In other words, not data as an output, but as the core input.

By way of example, our analysis of data from the Rwanda Utilities Regulatory Authority has enabled us to analyse mobile money trends in Rwanda from before and after the fee structure was changed. (These changes – implemented shortly before the COVID-19 lockdown began – resulted in the temporary zero-rating of fees on some mobile money transaction, particularly person-to-person transactions.) Consequently, we have been able to advise the government on appropriate fee structures using data about the impact of these changes on businesses and consumers. In a sense, government turns to data and says, “What should we do?”. The data assists in providing a view of behaviour. And by undertaking qualitative research to better understand the motivations for the behaviour you observe, you can start fine-tuning your response by, for example, testing to see how different transaction types or different markets respond to different fee structures.

Suddenly, governments don’t have to try a tool or policy intervention and wait for a year to see how people respond. They are able to see an almost real-time response to policy interventions. That’s an application of data at an entirely new level, and I think it is quite exciting. 

KK: What kind of skills are required to make good use of data to inform policy decisions?

HB: Four capabilities are required:

• Data discovery – which isn’t as obvious as it might seem, because governments often sit on datasets that they’re not even aware of, let alone analysing. Discovering that you have the data, and then processing that data so that it can be analysed in a manner that respects the privacy of customers and citizens, is the first skill required.
• Policy fluency – the ability to understand what questions you must ask and what questions you can ask given the data you have, or what data you need to answer those questions. Almost all policymakers can ask questions, but the ability to interpret the policy need, given the data, is sometimes lacking. This is not a data science question, it’s a capability that bridges the data with the policy needs of the policymakers. I think that is a skill governments will have to develop in the future, but which is currently quite rare.
• Data science skills – sophisticated data science skills are required in order to work with big data. This skill is already present among policymakers to some extent, but may need to be refined.
• Interpreting “data answers” into practical policy measures. Take our Rwanda work – analysing mobile money transaction data to understand the impact of changes in transaction fees. If you see a particular behaviour change following a price change, you still don’t know whether it was caused by the amount of the price change or the simple fact that a price change occurred. To interpret that result, the data alone is not going to help you. You need to add other knowledge that you have about the market or the environment; you may need to do some qualitative research. Similarly, in the private sector some banks have struggled with high churn rates: In some segments, more than 50% of people who sign up for a new account will have a dormant account within as little as two months. In exploring why that happens, some of the answers will come from the data – but some won’t.

Data enables you to get to the problem much faster than with any other tool, but it doesn’t necessarily give you the solution. You still require classic disciplines of qualitative research, a good understanding of your customer, a good understanding of your market context, and the experience and wisdom to know how clients traditionally experience or respond to what happened. We often think that technologies provide a definitive solution, but they don’t: Instead, they’re an important new tool that we can bring to bear on the problem. I think you still need interdisciplinary teams to solve problems.

Private companies, particularly big tech firms, have had to hone these skills over the past two decades, but for governments this represents a new skill. Data and data analysis on their own do not yet deliver a practical set of policy interventions, you need to go a bit further. Cenfri (one of the hosts of insight2impact) is entering into new projects that will allow us to hone our own skills in this regard.

KK: How can the development community and the public sector assist in catalysing the use of data for policy reform or business strategy?

HB: There are lots of new datasets becoming available, but compared to what you have in the developed world, access to data is still quite limited. The development community and donors can play a significant role in funding, and also in using their clout to assist governments in overcoming some of the inherent reticence to make data, or government databases, public and usable for the private sector. I think donors have a large role to play in data discovery and creating the regulatory legal frameworks – whether that’s through working directly with governments to make data available, or whether that is, for example, creating data hosting competitions. There’s a lot to be done when it comes to governments fully comprehending the use cases, and how data can add value to them. The policy value chain, in its entirety, is still a subject requiring funding. Whether it is building awareness among policymakers, doing pilot projects with government to help them develop their internal capabilities, or holding partnerships with other government entities (where some data is available only at a regional level, for example), there are many areas that require support, particularly on the African continent. 

KK: What role should government play in managing the outcomes of increased data usage and digitalisation?

HB: One of the initial hypotheses for insight2impact – and I think it’s been proven correct – is that a lot of the economic interchanges or economic dynamics in Africa were simply invisible because there was no way to capture them. The economy was almost entirely informal, and most of the people didn’t have bank accounts with records, let alone digital accounts. Suddenly, with millions of Africans using digital devices, you get access to information that didn’t exist previously – the digital economy represents a quantum leap in reducing information asymmetry. You have a lot of information that was inaccessible and invisible becoming visible, and much of that information is about the nature of the demand and supply in the economy. Just having that data visible will, by definition, spawn new business models and new products, as well as new distribution channels and new logistical solutions.

When you insert that sort of catalytic driver into economies like those in Africa, you’re going to see exponential changes. I think that’s very exciting, but it’s not going to be all good – when people connect suddenly to as powerful a medium as the internet, and they become visible to the internet, they also become targets. The internet already acts like the opposite of an intravenous drip; it sucks financial resources from people in ways that just weren’t possible before, whether that’s from online shopping, or fees on transactions and other services – payments and otherwise. Many people see the role of the government as a regulator that can counteract this process, and there’s a need for consumer protection. But there’s also going to be a huge demand for proactive communication around these risks and opportunities. Developed economies have had decades to familiarise themselves with the gradual creep of the digital economy. When it is unleashed with such rapidity on a population like Africa’s, the governments – because they have purer incentives (we hope) than the private sector – will have to step in with massive national education campaigns.

KK: Any final thoughts regarding the end of the insight2impact programme?

HB: From myself, and from both insight2impact’s host organisations, FinMark Trust and Cenfri, there is enormous gratitude for the privilege of having been able to work on a programme that in many ways was visionary and catalytic. It was hard to implement at the time because it was so exploratory, but for our organisations and, I think, for many of our partners, it allowed us to pivot to a different space. To have had that experience just before something as revolutionary as the current pandemic was really a privilege. For example, I cannot see a Cenfri without a solid data practice going forward. We’ve always done extensive work in the policy field, and I cannot see us comfortably doing that in the future without having recourse to data analysis as another tool to deploy, in order to give optimal policy advice to governments.

This article was first published on NextBillion as part of the Openi2i series an initiative of Cenfri and Finmark Trust. insight2impact (i2ifacility) was funded by Bill & Melinda Gates Foundation in partnership with Mastercard Foundation. The programme was established and driven by Cenfri and Finmark Trust.