Better lending decisions with alternative data

The lending market is exploding thanks to the emergence of trust scores – but it must be a controlled explosion…

Trust scores and the application of alternative data to make good lending decisions about thin file applicants is good news both for the financial inclusion movement and for lenders looking to grow – providing the technology is used to make better lending decisions.

Alternative data is big business

Alternative data hit the world’s headlines recently when it emerged that the Chinese government plans to break up Ant Group’s Alipay, the country’s largest digital payments platform. The government is forcing Alipay to spin off its loans business to create a new credit-scoring joint venture that will be partly state-owned. As part of the ‘deal’, Alipay will have to hand over access to the data it uses to make lending decisions. The motivation may be more about acquiring citizen data than anything else, but it’s a huge blow for the Ant Group whose CreditTech arm accounts for almost 40% of the group’s revenues. Lending is very big business indeed.

What’s good for business can be good for financial inclusion

The Ant Group is just one example (a very big one) of the potential of businesses to expand the financial ecosystem and embrace low-income individuals and micro, small, and medium enterprises (MSMEs) with little or no credit history. In the case of Ant, it’s the incorporation of the Alipay digital payments product into e-commerce and social media platforms that has proved so powerful, producing huge volumes of alternative data which it uses to create consumer trust scores.

It’s easy to understand why the international development community is excited about the potential of alternative data to help the estimated 2 billion adults around the world with no access to financial products. Alternative non-financial data opens up even more possibilities and the World Bank itself has expressed its support for its use in credit origination processes.

Lack of access to loans is also an acute problem for millions of MSMEs in developing countries. 40% of them have unmet financing needs according to the World Bank Group’s SME Finance Forum. Yet most of these, if properly assessed and managed, would be able to service the loans they need without any problem. Indeed, for lenders specialising in this segment, a low, single digit default rate is the norm.

Why does alternative data have such huge potential?

Traditionally, there are three fundamental questions a lender wants answers to before agreeing to lend money:

(1)         Is the borrower who they say they are?

(2)         Can they afford to make the repayments?

(3)         Can they be trusted to make those repayments?

For many people, giving an acceptable answer to question (1) is a problem. Staggering numbers of adults around the world do not officially exist in the sense that they have no official documentation regarding their birth. More than 2.5 billion people around the world have no passport, government ID card or birth certificate. They are excluded from accessing most financial and government services. It’s one of the main factors behind the unbanked status of 2 billion people worldwide.

According to UNICEF, the births of one in four of the world’s under-5s have never been recorded. That’s 17 million in Nigeria (population 201 million) alone. In the Philippines (population 108 million) it’s estimated that 5 million Pinoys’ births are unregistered. And although registration levels are generally increasing around the globe, this obstacle to financial inclusion will persist for a long time.

But one of the great benefits of alternative data is that it helps reduce the importance of (1) to a lender by compensating with answers to question (3) – trust. In other words, it matters much less who the customer is than how reliable they are.

Where does alternative data come from?

Most of the useful alternative data has arrived from technological advances – e-wallets, e-commerce and so on. But nothing has changed the lending landscape in areas such as South-east Asia and Africa more than the rise in mobile phone ownership. 

Way beyond the ability to communicate, smartphones have granted previously excluded people access to at least a corner of the financial ecosystem. Mobile money products such as e-wallets have given people previously unimaginable opportunities to participate. According to the IMF’s Financial Access Survey, there are already twice as many mobile money accounts as bank accounts in low-income economies. Having a mobile phone and paying a telecoms operator for the service can themselves be indicators of financial responsibility and consistency. And because mobile wallets leave a history of money added and spent, they too provide alternative data whose patterns can help paint a picture of an individual.

But this is just a starting place for alternative data which the International Committee on Credit Reporting (ICCR) defines as:

The collection and analysis of data on creditworthiness based on information which is readily available in digitized form but ‘alternative’ to conventional methods such as documented credit history.

The first wave of alternative data has been based on structured data which is widely (but not exclusively) available through smartphone data and derives from activities such as:

But alternative intelligence (AI) technologies also make it possible to draw conclusions about people’s trustworthiness based on the unstructured data sitting on people’s phones such as social media and internet use, emails, text, instant messaging, and even images.

Looked at individually, these sources of data may seem to produce a very fragmented picture compared to traditional credit scores, but combined they can paint quite a balanced picture of an individual:

Some trust score pioneers

The other side of alternative data

For every ‘good news’ story about the power of alternative data to financially empower people there seems to be a negative one – concerns about data privacy and consumer protection, accuracy of data, lack of transparency, and lack of opportunity for redress, for example.

The real horror stories begin when initially sound lending strategies begin to derail, and lenders start focusing on more loans rather than better ones. The resulting ‘continuous loop’ of loans is toxic not only to the consumer but to the financial institution involved.

In Kenya, for example, where Tala led the way and dozens (50 at the last count) of me-too apps followed, much of the industry seems to have abandoned the idea of making better lending decisions altogether. With APRs of 180% (Tala’s highest) and higher (500% is not unknown), it’s not surprising that one in 10 Kenyan adults has defaulted on a digital loan. Many find themselves in a debt trap, juggling loans by taking new ones to pay off old ones, and even being driven back into the arms of moneylenders. The Kenyan government says it will step in and regulate digital lenders but for millions of Kenyans the damage is already done.

One important lesson is this: whatever the benefits promised by alternative data and AI, a sustainable lending strategy is always a marathon, never a sprint; a strategy based on customers wanting to come back, not customers who can’t leave. This is a strategy by which significant numbers of micro-lenders in Kenya have managed to build profitable and sustainable businesses lending to MSME and retail customers by using credit decision models which emphasise the ability to repay smaller loans more frequently, coupled with business education for their customers.

Another important factor is almost hidden in plain sight. It may seem counterintuitive for digitalisation specialists such as Oradian to be pointing this out, but a loan approval process which is 100% automated risks lacking human insight and empathy.  

Building trust scores into your lending business

It seems that every month a new company emerges with a new way to calculate trust scores from alternative data. As AI develops, even more solutions will emerge. So, if you’re a lender looking to use them to help you make fast and sound lending decisions, you need a core lending platform which handles not just today’s alternative data tools, but tomorrow’s.

Linking between multiple (and potentially changing) partners in a process that is seamless to the customer requires agility, flexibility, security, and speed.  This is a combination that can only be provided through Open APIs and flexible loan origination and loan management workflows. Coupling flexible workflows with trust scores based on alternative data makes it possible to automate 90% or more of the lending process, resulting in safe and sustainable business decisions which can deliver rapid, constant growth balancing new and repeat business.

Oradian’s core lending system is called Instafin. It’s designed to connect with the ever-evolving financial ecosystem using Open APIs, making it quick and easy to connect with – for example – the alternative credit scoring services of today and tomorrow. Instafin is future-proof.