How machine learning can redefine lending

By Darpan Saini | 21 December 2016

The lending ecosystem has witnessed momentous changes in last five years, from fintechs disrupting the industry by leveraging technology and offering ease and speed in process, to evolution of stringent regulations post the 2008 meltdown.

Technology has played a significant role in the rapid evolution of the lending industry. One such technology, machine learning, is beginning to create new avenues in the lending market.

Machine learning, in simple terms, is an extension of artificial intelligence that enables computers or robots with the ability to learn, analysse and predict, using algorithms that iteratively learn from data. Machine learning therefore empowers the system to learn and adapt itself.

No more a concept of sci-fi movies, machine learning is already a part of our lives in more ways than obvious. The much talked about Google self-driving car is an extreme example of machine learning. At a more generic level, the customised feeds on our Facebook walls and the online recommendations and offers from Amazon and Netflix are also enabled by machine learning.

Though machine learning is not a novel concept, the influx of big data and data mining has given it a shot in the arm by integrating it in our day to day lives. Machine learning today is being implemented in various industries, from financial services, healthcare and retail to transportation, and multiple domains like accounting, audit, marketing and sales. Gartner identified it as a top ten strategic technology trend in 2016, with advances occurring rapidly.

Per McKinsey Quarterly report, June 2015, in Europe, more than a dozen banks have replaced older statistical-modelling approaches with machine learning techniques. By doing this, some of them have experienced a 10 per cent increase in sales of new products, 20 per cent saving in capital expenditures, 20 per cent increase in cash collections, and 20 per cent decline in churn.

Lending as an industry can achieve operational and strategic efficiency through the implementation of machine learning across its process and verticals.

Companies are already leveraging machine learning to expedite the lending process. Per an article by Jeremy Liew in American Banker, companies like Zestcash, Progreso Financiero, and Global Analytics are getting big, and fast with the implementation of machine learning and big data in their system. They make borrowing quick and easy, and pass the savings onto the consumers in form of lower rates.

Machine learning enables predictive modelling in credit scoring. Credit scoring is an important process in loan management. While the traditional credit score uses basic statistical tools to arrive at the result, machine learning involves data mining at a large scale by aggregating data through wider channels like Yelp scores, social media activity, and real-time shipping trends. This consecutively delivers a more accurate and meticulous portrait of creditworthiness.

Machine learning can also help in streamlining the lending process, eliminating errors and expediting the loan application approval process. According to a 2015 Cleveland Federal Reserve survey, traditionally, a bank takes three to six days to respond to a loan request and much longer in determining the creditworthiness and final acceptance of the loan. By eradicating the man-hours required in analysing the creditworthiness and the background of the borrower, machine learning reduces human work efforts in the underwriting and origination process.

A lender can create a ‘classifier’, a simple algorithm to approve a loan application. Using machine learning, the loan application classifier can learn from past applications, leverage current information in the application and predict future behaviour of the loan applicant.

Machine learning can also help in predicting bad loans and in on-going monitoring of loans. With publicly available Lending Club data, one can implement machine learning to analyse and predict the loans which have more propensity to default much in advance to take necessary steps for countering it. On the other hand, machine learning can also categorise non-traditional borrowers that are currently not being catered to, but should be considered, resulting in potential growth of business.

It seems to be apparent that more and more lending companies will enter the domain of machine learning and big data, which will make lending even more quick, accurate and easy. It is merely a matter of time before the technology companies and fintechs will integrate an organic layer of machine learning in their already dynamic platforms.

Darpan Saini, CTO and Co-Founder, Cloud Lending Solutions

Become a bobsguide member to access the following

1. Unrestricted access to bobsguide
2. Send a proposal request
3. Insights delivered daily to your inbox
4. Career development