Alternative data: A saving grace for African financial inclusion but a ‘perfect storm’ without controls

Mobile phone data has the potential to unlock credit and other financial services, boosting financial inclusion. But its use must be subject to careful consideration, according to participants in a webinar organised by AFIS.

By Oliver Nieburg

Mobile money has been instrumental in reaching the unbanked, particularly in East Africa with financial inclusion in Kenya reaching 83% in 2021.

This is unlocking data from mobile wallets and micro-behaviours on smartphones that fintechs can analyse with AI and machine learning to build profiles of unbanked individuals and small businesses, complementing traditional credit bureau scores.

Commercial bank partners can then offer credit and other financial services to previously untraceable individuals and businesses.

“It’s all about, how we can better use the data to formalise the informal, to score the unscorable and to bank the unbankable,” said Riadh Naouar, Manager of FIG Advisory Services, Africa, at the International Finance Corporation (IFC).

Speaking at a recent AFIS webinar, he said alternative data may help Africa fill a $331bn SME funding gap and better serve groups lacking credit histories and collateral ownership, such as women and farmers.

Data reliability

Alternative data is wide-ranging: from transaction data in mobile wallets to typing speed and whether contacts are saved with full names and proper grammar.

Potential also exists to leverage more digital data to inform credit decisions, such as Agtech data on farm size, crop diversification and access to inputs for farmers.

While offering promise to expand financial inclusion, some see dangers in extrapolating predictions about customer behaviour from alternative data.

“We’re thinking of it as potentially a silver bullet or a holy grail, but really it could be a perfect storm,” said Dr Jameelah Sharrieff-Ayedun, CEO, CreditRegistry, Nigeria’s largest credit bureau.

She warned that alternative data was vastly complex, and only consistent sources analysed over time, rather than a one-off data dump, could truly provide a picture of customer behaviour.

Prospective borrowers may simply adapt to be scored more favourably, she warned. “If I knew that social media was going to be a determining factor of my credit, I’m going to start pruning my list of contacts,” said Dr. Sharrieff-Ayedun.

Traditional banks optimistic, but wary

Most of Africa’s traditional banks are only just beginning to experiment.

Omar Balafrej, Group Director, Business Development Unit, Bank of Africa, said: “Alternative scoring is something quite new for us. We are modest about it and don’t hesitate to partner with fintechs that have the specialties and models.”

Bank of Africa has teamed with international fintech Simbrella to score individuals using behaviours on mobile wallets and will imminently launch its first pilot for nano consumer loans in Burkina Faso.

But the bank says alternative data is a gamble: “Alternative scoring is most of the time very, very risky, at least for us,” said Balafrej. For its Burkina Faso pilot, the bank will be relying only on banking data and mobile wallet behaviour without requiring guarantees or details about the borrower’s employer.

Yet the Casablanca-headquartered conglomerate still sees potential to add new tech partners with alternative scoring expertise to extend credit to SMEs, a priority segment for the bank.

Bank of Africa is currently partnering with IFC and fintech Manobi Africa to score agricultural cooperatives and small agribusiness in Senegal.

Shared telco datasets

Telcos are the holders of data and are themselves starting to offer financial products using alternative metrics.

MTN Uganda is using alternative data for its MoMoAdvance (overdraft facility for mobile wallets in partnership with NCBA Group), MoKash (short-term loans), and MoSente (a mobile lending service in partnership with Jumo).

Dennis Musinguzi, Acting Chief Product Officer, MTN Mobile Money Uganda, said that analysing alternative data on mobile wallets can be tricky.

“[Airtel Uganda] and us have struggled together, especially in this space of providing credit to our market because customers can just switch between sim cards, depending on where their debt lies.”

Mobile money users may spread incoming and outgoing funds across different mobile wallets making it challenging to interpret credit risk, he said.

Uganda’s two leading telcos are discussing developing a shared dataset with the regulators, allowing both companies to analyse mobile wallets effectively and see whether customers have debts with other providers.

Customer consent

Fintech Yabx – part of India’s $12bn Mahindra Group – has a credit scoring model based on mobile wallet data from MTN, Airtel and Orange that primarily provides a proxy of customer income.

This is helping to identify small-scale informal traders, who can be offered merchant loans via collaborative products from Yabx’s partner banks and telcos.

According to Eunice Ruguru Gatama, Director Africa, Yabx, companies need to not only be upfront that they are mining customer data but explain how that data will be used.

“I think the majority of the industry is very quick to ask the customer to consent, but then later on the customer is not sure exactly what they consented to.”

She said an unexpected and unwanted example could be a lender mining the contact details of a loan holder’s mother during credit scoring, and then contacting her to urge her son or daughter to pay back their debt.

Data privacy

Most African countries (36 of 54) have implemented or developed data privacy regulation.

“I believe the majority of them are good enough to protect the rights of data privacy for all of us,” said IFC’s Riadh Naouar.

He expects African policymakers – that have already shown agility to allow mobile money to thrive – will be able to adapt regulatory frameworks for an expanding data economy.

This will include keeping a check on the market power of big tech and social media companies which capture alternative data and are exploring an expansion into financial services.

He added that there should be regular audits for those licensed to gather data to ensure they have adequate governance and security.

Non-financial data: A grey area

Bank of Africa’s Omar Balafrej said that while financial data may come under the remit of central banks, governments should set criteria on how non-financial data, such as social media, can be used.

“This is the role of the governments to set rules on this. What can be used and what cannot be used, and what should be deleted after a certain period and not stay with that individual,” he said.

As non-financial data is a grey area, Bank of Africa is only using mobile wallet data in the retail market.

“We want to use more, but with our high standards of ethics, we are not ready to compromise any values on this,” said Balafrej.

The webinar panellists concurred that alternative credit scoring should not only be deployed to lure customers, but to offer affordable credit adapted to customer needs, ideally rewarding them with lower interest rates when they have started to repay loans consistently.

Watch the full webinar replay: