It’s no secret that online lenders are bringing efficiency to the borrowing market, providing expedited loan processes for personal loans, student debt refinancing, mortgages and small and medium business (SMB) loans and credit lines. Helping power this efficiency is big data technology to dynamically calculate loan terms as well as analyze risk factors of borrowers.
Among new data sets are alternatives to credit score rankings such as behavioral, social and psychometric evaluations. While this information is less fundamental in regard to setting loan terms, the data can help lenders better understand the risk and expected repayment characteristics to boost their acceptance rates. Due to the costs of acquiring new customers and potential profits from each loan, even a 10% increase in serviceable loans can make a big difference to lender bottom lines.
As such, while online lenders battle each other for market share and clients, alternative data providers are also competing with their analysis methods to market their solutions to lenders. In this Fintech Spotlight, we take a look at two alternative risk assessment methods and the startups providing them.
Popular in the UK and other parts of the world, psychometric evaluations are often used to analyze potential job hires to better understand their behaviors and likelihood to succeed in the
company. Among the questions are situational queries to evaluate how people tend to handle different tasks and life events. Based on the answers, employers can gauge how job applicants would be expected to deal with work in their company.
But can this information also be used to help lenders understand which borrowers are more or less likely to repay loans? According to Innovative Assessments co-founder and organizational psychologist Dr. Saul Fine, the answer is yes.
Speaking to Finance Magnates, Fine explained that research has shown that borrowers also have certain types of personal attributes that can be discovered in psychometric evaluations. As such, questions about a borrower’s attitudes and preferences regarding financial decisions can predict their future willingness to pay off debts.
Skeptics to this form of data for determining risk assessments though have questioned whether answers can be false, with borrowers ‘gaming’ the test and providing answers that are most likely to satisfy lenders. While this concern also effects job evaluations, it is more relevant to loan applications where only a few psychometric questions can be asked before creating too much friction in the loan process and driving away borrowers.
On this problem, Fine explained that part of creating psychometric tests is “formulating questions that are difficult to fake, while also developing algorithms that can detect faking”.
Overall, Fine indicated that while psychometric tests might not be used to determine how much to lend and what rates to offer, he views his risk scores as a complementary product that can be integrated with a lender’s existing risk evaluation to improve its results. Fine explained: “Primarily, the score is designed to help lenders reconsider approving certain applications that might otherwise have been declined, while also indicating potential risk among some typically approved applications.”
On the other side of the coin is behavioral science, or also referred to as psycho-linguistics analysis solutions. With this method, data firms analyze correlations between multiple factors of a potential borrower’s life. This can include social items from Facebook and Twitter, career presentation featured on their Linkedin profiles and mobile usage available from phone bills.
According to Maoz Batzia, co-founder of Credible, peer research conducted to evaluate behavioral science correlations proved that it can be used for calculating the potential risk habits of people. According to Batzia, by using this information, tests can be created to gauge items such as how likely a borrower will repay a loan or whether an entrepreneur will be careful with funds invested in his company.
Batzia explained that in the lending world, behavioral science analysis provides a solution for evaluating the more than one billion unbanked people around the world. As such, in addition to helping lenders better understand their users in more established Western countries, behavioral data can increase lending in regions with underdeveloped banking systems and can promote financial inclusion.
Which is better?
With regard to which is better, each has its faults. Psychometric tests have proven themselves over the years to be useful for employment evaluation, and have developed a real track record. However, even when creating quick tests that can be minutes instead of hours long, they do increase the duration of the loan process and increase friction.
On the other hand, behavioral science analysis can be conducted in the background while mining data from multiple sources, and provide risk scores in seconds. However, application of this information has had a much shorter lifespan of proving itself in real life lending cases. In addition, there is debate as to how reliable behavioral calculations are due to changing social habits evolving quickly in the world.
Regardless of which method has the upper hand, the bigger trend is that big data functions is the major driver of the online lending market. Whether the loan is from a fintech startup, multi-national traditional bank or established online lender, innovation in analyzing borrower data will be in demand for the foreseeable future.
(Image credit: Pat Loika- Flickr)
Fintech Spotlight is a new column on Finance Magnates devoted to reviewing innovative financial technology companies and sector trends.