announcements‎ > ‎

Community banking through peer credit scoring

posted Aug 22, 2012, 10:08 PM by Ignacio Mas

 [From Center for Financial Inclusion blog, 22 August 2012] 

Giving personal credit is about judging people. Banks have established information standards about customers, designed to make the judging process routine. How long have you lived at a given fixed address? What is your level of education and profession? How much salary do you earn? What is your history of past debts? etc.

The problem is that this standard information set is not at all standard for the informal majority in developing countries. People may not have a fixed address, they are likely to be jacks-of-all-trades, no one guarantees them a wage, and prior borrowing and repayments have left no trace.

There is much talk about using online social networks to create a new systematic profile for customers whose data is now invisible to banks (such as Lenddo). The deal seems to be: reveal more to the lender about what you do, how you spend your time, what your interests are, who your friends are, and it will be better able to judge you. With today’s computing machines, the information collected from people can be a lot more bitty and diverse, and the code will crunch it into a standard profile. Powerful stuff: now, pretty much any information can be thrown into the judging machine.

Dave Birch, a payments expert at Consult Hyperion, does well to question two premises behind all this. First, is what people put up on social networks sufficiently truthful so as to be relevant, starting with their names? The bet is that social networks are now so critical to how people run their lives and so intricate and public that no one can afford to misrepresent themselves any more. Gone are the days when you could confidently think that on the internet, nobody knows you are a dog. Second, will people really want to friend their banker, giving him a window into their intimacy? They may, but they’ll resent it and then they may come to regret it, because the other thing that has changed with the internet is that there is no more taking back of information once it has been put out there.

Social banking feels very new, but it is in fact old banking logic extended: it’s about the lender gathering relevant information to make better judgments. But how long really before the informal majority gets caught in banks’ digital information net? And how useful will their digital footprint –the traces of their online existence captured by Google, Facebook and such— be, anyway?

At the opposite extreme are peer-to-peer online lending platforms such as, where borrowers are the ones putting themselves and their information actively on display. Prospective borrowers seek to meet up with individual lenders in a digital lending marketplace, cutting out the bank from all information collecting and judging. Here the issues with truthfulness on the part of the borrower seem more acute, because it’s harder for individual lenders to check all that information put up by the borrower.

There is another way. Microcredit, at least in its pure early group lending form, was not about centralizing the information but about distributing the judging. The lender may never know enough about the borrower, but borrowers know plenty about each other. So put incentives on borrowers to select and monitor each other. Bingo! A microlending revolution was born.

Surely this principle can be applied much more deliberately and forcefully in the digital age, where one’s peers need not be restricted to those within walking distance. The internet is about information decentralization, not necessarily in the sense of where it is physically held (those Google servers are looming rather large) but in the sense of who put it out there and how it is being used.

How might this work? I described such a scheme in this paper where I try to imagine what a totally mobile bank might look like. I visualize banks assigning a dual credit score to their customers: a personal credit score relating to the customer’s personal data and own debt history, plus a social credit score relating to how well the customer has judged her peers’ performance with debt. Customers seeking debt beyond what their personal credit score can justify can ask other customers to vouch for their creditworthiness. It’s not a guarantee: there is no financial penalty for vouching for someone who ends up not repaying, except that the bank will listen to that person’s opinion of others a little bit less — her social credit score will be reduced. The opinion of a client who keeps endorsing bad fellow clients will be increasingly ignored by the bank, as her social credit score eventually hits the floor at zero (but, again, her personal credit score and ability to get credit herself will not be affected).

Why risk that by vouchsafing others? To help your trusted friends in a way that is in the end costless to you. But also because some people may aspire to be trusted by the bank, to play surrogate banker in their community, and the only way to increase you social credit score is to take calculated risks on it. I’d expect that the bank would eventually zero in on certain individuals in each community who have developed a good track-record as referees. A new way will open up to build profile in your community: building up a profile with the bankers.

Banks can embark on the path of peer credit scoring incrementally, relying on the private credit score more at first, and weighing the social one more over time as people’s peer recommendation track record builds up. This system of vouching can be easily built into simple mobile banking platforms running on SMS or similar. Is there any reason why most banks in emerging markets couldn’t begin experimenting and building trust networks between their customers today?

Unlike peer-to-peer lending, this system maintains a key role for the financial institution, which is desirable because we expect the institution also to package such credit with savings and other financial services people need.

Unlike social media lending, this approach doesn’t require intrusive information gathering by the bank. No big number-crunching machines, just leveraging the knowledge of your own customers. And in any case, those with a banking deficit are also likely to have an internet access deficit.

Unlike traditional group microcredit, you can tap the support of friends and relatives anywhere, without the need for group meetings.