The link between your cell phone and your creditworthiness: fact, or fiction?

By David E. Falconer

Manager, Sales and Business Development, Creditinfo Guyana

 

Almost a century ago, an American named John Pierpoint Morgan, or J P Morgan who is more widely known as the most powerful financier in the world, when asked by a House committee if commercial credit was based primarily on money or property, responded instead to everyone’s surprise that in fact it was based on character.

His words are so portentous and important that I believe it is useful to replicate them for your benefit: “The first thing is character… before money or property or anything else. Money cannot buy it… because a man I do not trust could not get money from me on all the bonds in Christendom.”

This central issue of character and trust is as important today as it was so many years ago to old JP Morgan. Banks and other lenders look to an applicant’s past records and behaviour to derive some indication of how he is likely to behave in the future. Credit markets today function on a foundation of trust; but this is trust that is built firmly on the knowledge of past behaviour and an expectation that this behaviour will persevere in the future.

Any other system where blind trust and gut instinct feature prominently is the stuff of bad movies and fictional tales and will surely lead down the road to perdition. Will it not?

This concept has worked particularly well for credit bureaus which collect vast amounts of past credit and credit-related information, subject it to algorithms that assess trends and patterns, and then produce credit reports and scores that essentially reduce these mountains of data to a more meaningful assessment that lenders can use.

Other than that, when you walk into a bank and ask for a loan, your banker will pull up your financial history – your bank accounts, your credit score, your assets and liabilities. This is how the bank determines if you are a good, bad, or acceptable risk and how they determine interest rates, collateral requirements and the like.

But, herein lies the problem. Available data shows that many persons and small enterprises in developing countries (like Guyana, for instance) lack this formal financial history. Without this data, you may ask, how can potential lenders evaluate the character of their applicants and derive appropriate levels of trust? Isn’t it then difficult for banks to allocate capital, and for potential borrowers to obtain loans?

Lenders then, may have little formal information on some potential borrowers. This is seen as particularly problematic, as banks who lend to the unbanked, or to small or informal businesses may have little recourse if they were to default. Even if they were to rely on the legal system, it is usually too costly to recover small loans that have gone bad. It is clear that there has to be some way to properly evaluate creditworthiness prior to disbursement.

The answer is way, way out there. Most everyone these days have mobile phones, don’t they? Apparently, even the pre-paid versions generate massive amounts of useful information about their user’s behaviour.

It transpires then that behavioural signatures inherent in mobile phone data can predict default with unusual accuracy, even approaching that of conventional credit scoring models that rely on financial data!

So then, although the unbanked lack the formal records needed for traditional scoring methods, many have yet managed after all to maintain a rich history with a formal institution over an extended period of time – their individual and collective mobile phone histories as recorded with their service provider.

As of 2011, there were 4.5 billion mobile phone accounts in developing countries around the world (International Telecommunication Union of the United Nations, 2011) and even pre-paid phones give up a valuable store of data on individual behaviour and social networks. If indicators derived from this data are somehow predictive of creditworthiness, then banks can identify qualified (and therefore profitable) borrowers they could not previously assess.

But why the hue and cry about these unbanked persons? Why not just leave them be? Consider this while I tell you that World Bank President Jim Yong Kim launched a campaign in April 2015 to achieve universal financial access by 2020, and consider also that about 2 billion people across the world currently lack access to a transaction account.

Last year, a study done by Daniel Björkegren and Darrell Grissen supported by the Stanford Institute for Economic Policy Research matched call records to loan outcomes for a sample of borrowers in a Caribbean country. They found that about 1 in 8 of these loans defaulted, and a look at mobile phone data in the 90 days leading up to when persons received the loans showed that there were distinct patterns that linked the way people used their phones and the likelihood of them repaying the loan. These patterns proved remarkably accurate in predicting who was going to default, and who was not.

Some of the behavioural indicators that are credibly predictive of responsible loan repayment are as follows:

 

  1. A responsible borrower may keep his mobile phone credit at a minimum threshold so that he can have its use in case of an emergency, while a less responsible person may allow it to run out and depend on others to call him – this was found to be predictive of a tendency to default.
  2. Persons who charge the same amount of airtime on the same day every week are better credit risks than those who purchase a large amount, then let their accounts remain empty for long periods.
  3. An individual whose calls are returned may have stronger social connections that may support their pursuit of entrepreneurial opportunities.
  4. A person who has lots of contacts is a stronger credit risk than those who do not. In plain speak, if you happen to receive funding to start a business; you are more likely to succeed because you have demonstrably more friends who will support you.
  5. Getting more texts than you send correlates with higher creditworthiness, another sign of a large network of friends/contacts.
  6. Surprisingly, data shows that persons who gamble are more likely to pay off their loans. Perhaps because they are more accustomed to debt and sensitive to the need for responsible payment?
  7. Apparently, the more quickly you drain your battery, the less creditworthy you are. Clearly, if you have lots of time to make calls, send messages and use social media, you may not be a responsible person or hard worker.
  8. Data shows that persons who won’t go to diverse locations and instead choose to stay within familiar confines are less dependable borrowers than those who do.
  9. When phones stay in the same place every day, that’s usually a good sign that the user is at work. And average usage (or lack thereof) during working hours for certain classes of consumers can point to commitment and responsibility – both reliable indicators of a general tendency to repay loans.
  10. Persons who make more calls in the evening, which is off peak hours are considered low risk because of their price sensitivity.
  11. In developing market economies, mobile phones are increasingly serving as ledgers for the movement of money. Here in Guyana, GTT’s Mobile Money as well as the rise in popularity of online banking (not everyone has a computer but almost everyone has a smartphone) is proof positive that, in our own way we are on right on track with popular global trends. So, monitoring phone records becomes a simple substitute for examining a bank account, and sometimes far more efficiently since one can conceivably track accounts for multiple banks from the records of one phone.

First Access, a start-up company based in New York, mines data stored on borrowers’ mobile phones in Tanzania and Kenya and uses this data to recommend a loan amount to banks and microfinance lenders within 90 seconds. Many of the observable trends that drive these recommendations are listed above.

These proprietary algorithms can cut down clients’ underwriting costs by more than 92 per cent, according to co-founder and CEO Nicole Van Der Tuin.

The company’s backers include Capital One Financial Services co-founder Nigel Morris and The Social Entrepreneurs’ Fund – Bill Ackman’s venture capital fund investing in businesses that promote financial inclusion.

Of course, many of these studies have to be replicated for the point to be properly proven and we may still be some distance away from that. However, one cannot deny the simple pleasure of knowing that the ever widening scope of useable data available seem to point to the possibility of greater access to credit, faster approvals, lower rates of interest for qualified borrowers, lower default rates and more importantly – financial inclusion for a far greater number of potential borrowers than ever before.

And credit bureaus are right there to help. After all, who knows how to handle vast amounts of data better than we do?

Good news for banks and other lenders, and better news for the borrowers out there.

Everybody wins.