Digital banking: Credit scoring goes social

Rene Griemens, CFO at Kreditech, says 'there is a tremendous opportunity to reach the unbanked customers in the emerging middle-class in the emerging markets'

Will banks be disintermediated from lending? It’s not unthinkable. An increasing number of independent lenders are gaining traction in the loans industry, thanks to the use of technology to analyse credit risk. They lend where banks will not.

Lenddo was set up in 2011 and uses data from LinkedIn, Twitter, Google, Yahoo, Live and Facebook among other sites to form a profile of the borrower. Over 100,000 credit decisions have already been made through Lenddo.

“It’s much like JPMorgan himself looked at banking - where lending was based on the ‘character of the man’. We discovered that how you connected to others and who you are connected to will determine how you behave as a borrower,” says Jeff Stewart, co-founder of Lenddo.

Kabbage, headquartered in Atlanta, Georgia, makes loans to small businesses in the US and the UK. Since 2010 it has lent $300 million to 30,000 customers. Like Lenddo it uses some social-network data to complement third party credit scores, and it also improves the scoring by using data provided by the sales platforms its lenders use – such as UPS, eBay, Amazon, Etsy.

Says co-founder and COO, Kathryn Petralia: “Banks have always found it hard to profitably provide loans of less than $100,000 due to the highly manual nature of their lending process. They often require personal guarantees, and business owners are afraid to put their house on the line. Most banks can’t lend to business owners with a FICO score less than 700, but these borrowers are just as creditworthy - many of them are simply using personal credit to run their business, which artificially deflates their score. We supplement that data with our own collection of data to make credit decisions.”

Kabbage uses buyer feedback ratings, selling history, turnover, accounting data, bank account information and other metrics to approve cash advances.

Kabbage also focuses on providing loans within minutes, adding funds to borrowers’ PayPal and checking accounts. Petralia says: “Our automated platform enables us to offer funds immediately, which is difficult for banks and other financial institutions to do with a manual process.”

Local partners

The use of social networks in providing insight into a borrower’s credit risk is particularly useful in emerging markets. There the middle class is growing at a fast pace and often requires loans, but individuals lack a credit history. Lenddo provides banks with approved borrowers in Mexico, the Philippines and Colombia. Stewart says Lenddo hopes to reach 1 billion people in emerging markets within the next 10 years. “The plan is to find local partners and to keep adding more countries,” he says.

Indeed emerging markets have been a key focus for the new lenders. Kreditech, which launched last year is already in seven markets. In September it will open in the Dominican Republic, towards the end of the year in Romania, and it is looking at Brazil as well as others for 2015.

We discovered that how you connected to others and who you are connected to will determine how you behave as a borrower Jeff Stewart

“Eighty percent of our markets are between G8 and developing countries,” says Rene Griemens, CFO at Kreditech. “There is a tremendous opportunity to reach the unbanked customers in the emerging middle class in the emerging markets. We are also looking at African markets, which is really an economy-building function. It is very difficult for a high street bank to deliver what we offer, as loan amounts are so small. We are in discussion with one of the world development organizations around funding for expansion into emerging markets.”

Stewart says these lending services are having a social impact in emerging markets. “Often individuals with good references and a good profile are borrowing for their family’s education, or to help those less fortunate, as well as growing their own businesses.”

Lenddo also provides advice on financial health to its customers such as how to save or how to budget. Says Stewart: “We offer tips in Twitter-sized digestible chunks; interestingly how individuals engage with that information is also predictive of whether they will pay back the loans.”

In April Lenddo launched a Visa credit card with ScotiaBank’s Colombian subsidiary Banco Colpatria that is dependent on a consumer’s reputation on social networks. “We received 2,000 applications within the first month of launch of the co-branded credit card,” says Dan Gertsacov, chief executive for the Americas at Lenddo, “more than doubling Lenddo and the bank’s estimate and demonstrating the value of social networks to provide access to credit to a segment that had previously been ignored by traditional data analysis.”

Santiago Perdomo Maldonado, CEO of Banco Colpatria, says the use of social network data for banks in his country offers real value. “We’re facing a very important challenge helping people to access financial services, and this is an opportunity to really help Colombians to improve their life quality through financial services.”

Kreditech’s impact in emerging markets might be even greater than its peers as it does not rely on third-party credit scores. To test the credit-scoring technology it had developed, the firm started lending, and it was so successful in building the lending operations that it decided to keep doing that rather than become a credit-scoring vendor. Now it has scored more than 1 million individuals and will probably issue more than $100 million in new loans in 2014

“Most players tend to use the traditional credit-bureau scores and then improve on that using other data and social-network data,” says Griemens. “That means they cannot target customers that do not have a credit score. According to the World Bank, that is some 5 billion to 6 billion people.”

Kreditech instead has developed its own scoring model that uses public and private sources about the customer. Originally the firm started in payday lending or microlending; Griemens says that is no longer a core market.

“If you cannot rely on your credit scoring, then you need the high interest rates to buffer it – you usually lose on the first loan and recover the losses on a second. But we only lend to about 10% to 20% of applicants and have fewer defaults – it’s about 10% in Russia and Mexico, and Poland runs around 7% - better than the banks.”

Kreditech has average growth rates of 80% a quarter in both loan volumes and revenues. “About $30 million in cash revenues after defaults – more under US GAAP,” says Griemens.

Running scared

Should the high street banks be running scared of the new emerging lenders? Griemens says they are likely to be driven out of business unless they have the courage to overhaul their models.

“The incumbents are never as good as a dynamic market entrant because of the legacy technologies that they are beholden to,” he says. “Consultants come in and advise them, and systems and processes are redesigned and upgraded, but really they are working with processes established decades ago. Banks will be the next wave of incumbents to be driven to the wall. They will find it hard to close their branches, yet they will have to.”

Griemens adds: “At the moment banks have a low cost of capital because of their deposit base, but as nontraditional competitors find their cost of capital diminishing, those competitors will be a serious threat. We are borrowing cheaper every year, and next year hope to introduce deposits where we can pay the interest rate to match our loans – without the fixed costs that banks with branches have.”

Kabbage’s Petralia says that banks will most likely team up with alternative lenders. “There is innovation in banks, but culturally they are slow to move. They are struggling with increasingly complex regulation. However, banks recognize the huge opportunity in small-business lending, so we will likely see more partnerships in the future.”

Cameron Jones, security intelligence director at business analytics software provider SAS, says they should team up with them, not just to provide loans but to help battle fraud.

Pooled data from across banks (the so-called shift to big data) combined with social network data could help banks save millions of dollars in both fraud and regulatory fines, he says, pointing to one banking client that saved $45,000 in fraud costs in one week, and $2.3 million in the first year by pooling data from across the bank’s silos.

Jones says banks can see the importance of looking at their client information holistically. “There is so much data and institutions are overwhelmed, but the first task is to pool the data from across the silos. It doesn’t require any major organizational change for the firm or the added costs of additional hires. But it can save a lot of money if properly initiated.”

He points to simple red flags such as variations of names and addresses of the same individual showing up in the mortgage department data, credit card department data and deposit account data. “You can also get a better picture about a client with regards to credit risk when you can see the same customer has four credit cards, three mortgages and several accounts with you,” says Jones.

He says that banks are grappling with the use of social media in providing data and could look to more innovative emerging firms for guidance. One jewel in the crown could be a GPS-type system.

“The biggest point around social media networks is the location data that is built in,” says Jones. “It’s very easy to see where your customer is versus where their financial activity is taking place and prevent fraud or regulatory violation.”

He points to social-network location data as being useful in ensuring clients are not evading US taxes – a reference to Credit Suisse’s $2.6 billion settlement in May this year over accusations of helping US clients evade US taxes.