It’s a Catch 22: to get financial credit, you need a credit history; to get a credit history, someone has to give you credit. And in many developing markets, that’s unlikely to be a bank. Not for the first time, the mobile phone is well placed to come to the rescue – but this time, it needs a little data science help.
The mobile phone delivered connectivity to the unconnected. It brought wireless communication, email, commerce and the internet to parts of the world that had struggled to deliver a traditional wired infrastructure for a mass consumer audience.
And now, the mobile phone is doing the same with banking. At the end of 2016, there were more than half a billion mobile money accounts registered worldwide – including in some two-thirds of those countries with low or middle income economies. More than any other innovation or development, the smartphone is delivering financial services to the unbanked and is able to reach a population living outside the infrastructure of the traditional banking industry.
Today, the banking services available to those 500 million customers are usually fairly basic. They can make and receive payments based on the existing balance in their accounts, but for many among the things they can’t do is build a financial identity and get access to credit and upstream financial services.
Mobile money has successfully expanded access to a range of financial inclusion services for the previously unbanked. However, the reason it has not, so far, been able to unlock credit facilities for its customers, is because of its reliance on the same credit-scoring techniques that exist in traditional banking.
Traditional banks always want to check an established credit history, and they like to authenticate identities and know something of the background of the people that are taking out the money loan. In contrast, the vast majority of the 3.7 billion pre-pay mobile accounts in the world are anonymous; which means that with a pre-pay mobile banking customer, it is usually the case that all of that traditional banking background data is missing.
Behaviour holds the key
Nevertheless, financial services providers are continually looking for ways to expand their customer bases – and the millions of mobile money customers are an obvious opportunity. Providing access to small amounts of micro-credit is a natural gateway to more formal credit schemes, and data science can be used to build behavioural data that can expand access to financial services.
In many developing markets, the levels of financial exclusion are particularly high among women even though they are often the primary money managers and savers. In Pakistan for example, some 39 per cent of adult women are responsible for paying the bills and saving for the family’s needs and yet earn no income themselves.
On paper, and certainly in the world of traditional banking, this makes them a bad loan risk. They have no way of showing they are reliable borrowers able to pay loans back with discipline. What’s even more frustrating is that the lenders know these people exist but, without reliable data, they find the task of identifying them almost impossible.
It’s obvious that the providers need a better way to target the individuals that present a minimal credit risk, because the methods and data used in the developed world will fail in developing markets. In fact, according to the World Bank there are some 1.5 billion people in the world without any formal ID to prove who they are or to confirm their reliability.
Building a better picture
With smartphone users in emerging markets expected to top three billion by 2020, the mobile operator has unrivalled access to the unbanked and the undocumented. What’s more, the operators also have access to vast amounts of behavioural data about those individuals.
Partnerships between the financial companies and the operator community have developed over the past ten years or more, and have helped to create this mobile money market of more than 500 million users. But despite that remarkable progress – the application of the data science techniques we have been pioneering could yet turn out to be the most significant game-changer in terms of financial services.
We partner with mobile operators to deploy ‘Identity Scoring’. It begins by offering small credit extensions to pre-pay consumers and then applying data science and machine learning techniques to their payback profile. Further analysis of their behaviour against subsequent mobile transactions, rewards or incentive schemes adds to the data points we use to build a fuller financial profile of the customer.
Based on the sum of these interactions we can then establish personalized lending criteria for individual customers and walk them up a pathway to expand their access to more advanced personalized financial services. And at each step of the process, the application of data science is minimising or eliminating the risk for the providers.
In our experience of applying these techniques, in some 23 different markets and economies, we have seen consistently high repayment rates across the board regardless of the strength of the local economy. More importantly, Identity Scoring has enabled us to expand access to financial services to tens of millions of creditworthy individuals, all of whom would be outside the scope of traditional credit scoring methods.
Unlocking the potential
Despite the undoubted successes so far, there is still enormous potential for mobile operators and financial service providers to collaborate and address the challenge of financial inclusion in the developing markets. The potential still dwarfs the achievements to date.
However, traditional banking approaches will not deliver the required results. Only big data analysis, and the appliance of data science and machine learning will unlock the vast untapped potential. As ever in finance, it is all about the numbers – but in this case, it’s all about the science too.