Machine Learning: More than Just a Buzzword for FIs?
May 2019

If you live anywhere except under a rock, you’ve probably heard the term, “Machine Learning,” thrown around a lot lately. Well, there’s a reason! The world is accelerating towards a reality where it’s hard to go an hour without being impacted by machine learning. It’s in your phones, cars, TVs, package deliveries, and maybe even your thermostat for goodness sake.

So, what is it? In short, machine learning enables systems to use data to draw predictive or inferential conclusions without explicit human interaction. For example, suppose a programmer wants to build a robot with four legs that can walk. One (non-machine-learning) approach would be to write code that explicitly tells the robot what series of motions are required to walk. However, an approach that uses machine learning might instead involve code that tells the robot to make random motions, noting the motions that lead to the desired forward motion. Over time, as millions of motion combinations are tried and learned from, the robot will awkwardly progress towards ideal four-legged walking.

Now, most financial institutions aren’t in the business of making computers walk. But boy, do they have data. And where there’s data, there are endless machine learning possibilities. I’m going to list three that, when used properly, could have a huge impact on your business.

1. Credit Risk: This a complex facet of lending that has (appropriately) received a wave of attention since 2008. With CECL, DFAST, CCAR, and other regulatory responses to the financial crisis, FIs are required to step up their predictive game. In this setting where data is available and predictions are necessary, there are a number of machine learning algorithms and models that might fit the bill. Visible Equity uses a Cox competing risk model that models the monthly probabilities of a loan defaulting or prepaying based on loan behavior and characteristics.

2. Fair Lending: Do your lending practices lead to disparate impact? Can you prove it? At one point or another, you may be required to answer these questions. Machine learning can assist you in doing so. Visible Equity’s Fair Lending module employs various machine learning algorithms and statistical tests that effectively find instances (or lack) of disparate impact.

3. Marketing: As you seek to expand your business, you can imagine how useful it would be to know which of your customers are on the market for your various offerings. There are techniques within machine learning that have amazing power to classify the optimal targets for your campaigns by observing behaviors and attributes in your customer base.

These are just three of many possible applications of machine learning for financial institutions. The industry is undoubtedly headed in this direction, and the benefits will be huge! If you haven’t yet, give us a call and join us for the ride.

Related Blog Posts
Further Your Education