October 24, 2016
FinTech is responsible for a multitude of significant achievements in the financial services industry including the transition from a closed-loop ecosystem to an environment that encourages open innovation, collaborative efforts and partnerships between institutional players and startups.
Aside from opening minds (and pockets) of institutional players, there has been another important transformation that happened under the influence of technology companies – the way financial institutions look at customers. More particularly, the way the creditworthiness of customers is assessed.
FICO has traditionally been the lens, through which financial institutions have been looking at the customer and, in fact, still is the most important factor in weighing relationships between banks and their customers. FICO has a relatively clear structure and breaks down into five categories: 35% – payment history, 30% – amounts owed, 15% – length of credit history, 10% – new credit, and 10% – credit mix.
Aside from being heavily reliant on financial behavior, the traditional scoring tool is also working in favor of a particular part of the population – eligible, banked working class with medium to high income. FICO may be a good tool in that sense as it appropriately unlocks opportunities for those people, but at the same time, it locks out unbanked part of the population or the ones with illegitimate history with financial institutions due to low income and other socio-economic hallmarks.
If one was to carefully assess the content of each element of FICO, he/she would be surprised how heavily everything revolves around various bank accounts, amounts of funds on accounts and transaction history: amounts owed on various accounts, loans, mortgages, utilization ratio, the number of accounts, etc. Unbanked population, hence, is excluded and cannot even shed the barriers to getting a fair assessment because they cannot have a bank account in the first place.
In fact, some estimates suggest that around 22% of the US population does not have a credit card. But, as we have mentioned, in order to have a credit score, you need to have some form of credit. Credit cards have played a crucial role in building or destroying credit score. In absolute numbers, other editions suggest that ~53 million people are currently not scored because they do not have active accounts that report to a bureau.
As Dan Macklin, SoFi Co-founder and Vice President of Community & Member Success, has commented upon alienation of SoFi scoring engine from FICO, Credit scores don’t provide an accurate picture for financially responsible professionals with a strong employment history and monthly free cash flow. These scores tend to be inaccurate, hard to dispute and even harder to pin down, with the onus falling to consumers to monitor multiple databases. We’re more interested in a comprehensive and forward-looking approach to assessing an applicant’s financial wellness.
Listing property, tax, deed records, checking and debit account management, payday lending information, address stability and club subscriptions being among the sources for alternative data. Mike Mondelli, SVP of TransUnion Alternative Data Services has emphasized, These alternative data sources have proven to accurately score more than 90% of applicants who otherwise would be returned as no-hit or thin-file by traditional models.
Alternative data provides a better lens with which to evaluate all consumers, giving lenders who can score them a competitive advantage, he added.
In recent years, technology companies have been discovering other alternative sources of data to determine creditworthiness, among which are Web search history, phone usage, social media and more. We have been strongly emphasizing the importance and inclusive role of alternative sources of data for creditworthiness assessment and even sensory technologies and their potential role in underwriting processes.
Modern sophisticated analytical algorithms built by tech startups that are able to extract necessary flags out of massive unstructured search data can probably become the next big thing across industries. So far, Chinese search giant Baidu and Californian big data underwriting company ZestFinance are on the path of using Web search history to extend microloans to China’s underbanked.
Baidu will be using ZestFinance’s underwriting technology to determine the creditworthiness of its users. Fortune cited a perfect example, explaining the determination of employment status. If an adult user is searching for video games in the middle of the day, he/she probably doesn’t have a job. In addition, if his/her previous history did not contain searches inherent to students, that person is probably not a student. Doubtfully, an unemployed adult – who is not a student – would represent the most creditworthy borrower as there is no indication of the ability to repay the loan.
Silicon Valley satellites like Affirm, LendUp, ZestFinance and others use data from sources such as social media, online behavior and data brokers to determine the creditworthiness of tens of thousands of US consumers who don’t have access to loans.
According to the WSJ, Visa has built mobile payment applications in Rwanda and is working with IBM to use records of retail transactions or remittances to create a surrogate credit score. Chinese e-commerce giant Alibaba Group Holding Ltd. was also reported to launch a credit-scoring program that uses the company’s own trove of transaction data to assess risk.
Smartphone usage is a whole another subject. Experts believe that smartphones can dramatically reduce the cost of lending because the apps they run generate huge amounts of data—texts, emails, GPS coordinates, social-media posts, retail receipts, and so on—indicating thousands of subtle patterns of behavior that correlate with repayment or default. The WSJ brought up examples of data pieces like the frequency of charging, the number of incoming messages in comparison to outgoing, the number of miles traveled in a given day or how users enter contacts into their phone—the decision to add the last name correlates with creditworthiness—can bear on a decision to extend credit.
Social media has been recognized by Wharton as an important data source for credit scoring back in 2014, although the practice of judging a stranger based on his/her social environment is not really new. One of the core ideas is that who you know matters.
As Wharton professionals suggest, if you’re a good person you must have some good connections with people around you, you must have a certain number of friends and you must have had this account for a certain period of time. Technology companies focusing on alternative lending and alternative credit scoring can gather so much more information about a person using social media than by looking at their financial data.
Companies like Lenddo, FriendlyScore, ModernLend are creating alternative ways to indicate creditworthiness. The information contained about a person in social networks can provide some sort of verification that the person exists at all and who that person is.
There are quite a few companies that are undermining traditional way of assessing credit risk and transforming the way financial institutions look at the customer. A variety of smaller Silicon Valley-backed companies has been working on building affordable and transparent alternatives to old-school credit cards and another set of companies proved that there is much more to assessing reliability than FICO.
What do those companies take into account? Is their approach different? Among the interesting examples is Shared Lending, a platform where people create their unique CORE Profiles. Their algorithm has no connection or correlation with FICO. Their model analyzes responses to five human characteristics: Productivity, Resilience, Finance, Health and Education.
Another example is ModernLend, an online lender that uses alternative metrics to identify and lend to international citizens who cannot access financial products due to lack of SSN or FICO score. The company’s first product is the ModernLend Debit Card that helps customers build credit histories in six months and access benefits from 300,000 vendors nationwide, all without requiring an SSN.
Tala (formerly InVenture) is also an interesting case, brought up by the Wall Street Journal. The company looked at phone usage logs to find out that users who wait until after 10 PM to make calls—when rates are lower—are lower-risk borrowers. Another data-focused company, Branch, found that, although counterintuitively, users who are known to be gamblers—something the app would find out by scanning messages or payments to a gambling company—are more likely to repay a loan than non-gamblers.
But, probably, the most forward-thinking case is the one represented by Aire, which has an alternate credit scoring model that uses data provided by users to develop a credit score. If the data can’t be validated, it doesn’t generate a credit score. The company allows people to build up their own credit profile and control the entire process. Moreover, all data is collected with user consent, unlike traditional credit-scoring companies that use a black-box method without any transparency.
It will certainly take time, but businesses and international financial institutions will follow through with advanced ways of assessing customers' financial reliability. There are several reasons why an alternative way of assessing someone’s creditworthiness will take hold:
The opportunity to ‘shed light’ on an increasing number of people around the world will unlock new markets for businesses and financial institutions as well as provide previously excluded parts of the population with a chance to build household resilience and give greater chance to survive various hardships.