Archive

From Our Briefings Newsletter

Published on20 NOV 2017

The prospect of higher interest rates has some investors worried about rising credit losses. But lenders today are using increasingly sophisticated tools that leverage big data and machine learning algorithms to better assess the riskiness of potential borrowers.  We sat down with David Stark, head of Risk Management for Marcus by Goldman Sachs, the firm’s recently launched consumer lending business, who described the changing landscape for lenders and borrowers.

The article below is from our BRIEFINGS newsletter of 20 November 2017:

Briefly . . . on the Evolution of Consumer Credit Analysis

David, given your experience in the consumer credit business, how has the process for evaluating potential borrowers changed?

David Stark: For years, lenders used to individually review borrowers’ applications in a process known as “judgmental lending,” so-called because the decision to grant credit was based on the lender’s judgment and experience. It was a difficult process to get right all the time in part because the process was subject to idiosyncratic and subjective factors. A lender who was having a bad day, for example, might be more inclined to reject potential applicants. The advent of credit scoring helped to systematize the evaluation process by analyzing how people repaid prior loans, how many loans they had already taken and how many times they had applied for credit, among other factors. All of this data was housed in a credit report and credit scoring boiled down that person’s creditworthiness to a single number — a credit score. Importantly, credit scoring brought a more objective lens to the process and over time, the models grew more sophisticated to incorporate more data.

There have been broader concerns that rising interest rates could lead to more credit losses and that this may be a bad time in general to expand into consumer lending. How prepared is the industry — and Goldman Sachs — to respond to a potential change in the credit environment?

DS: We’re very cognizant that the credit cycle won’t always remain as benign as it has been. But I would note that our management team has lived through multiple credit cycles and we have lots of learnings from those periods. Risk management and credit analyses are key pillars of this business. We’re not chasing growth for the sake of growth before the credit cycle turns — we’re building this business for the long-term. 

While we may be late in the cycle for consumer credit, the industry has gotten a lot better at assessing credit risks. Not only do companies have access to sophisticated analytical tools, but what’s also changed is the availability of data. It used to be that credit algorithms were based solely on a person’s data on file at the credit bureaus. But lenders are incorporating data sources outside of traditional credit reports, such as the applicant’s employment history, in making their decisions. When it comes to the Marcus business, because we’ve built our technology from scratch — using software through open-source platforms — we’re able to easily bring in new data sources. The more data you bring in, the more you can understand the nuances of behavior.

This sounds like machines are being used to a greater extent to make lending decisions. To what extent are people still involved?

DS: Despite the latest analytical tools, there is no substitute for actual lending experience so it’s important to build that experience into the lending process both at a tactical and strategic level. It’s also critically important to have an overarching risk framework in place. An individual analyst might have as much data as he or she needs to make a decision but may not have the experience to put that data in the context of the company’s total loan portfolio.

 

 

 

The data provided above is for information purposes only and should not be construed as investment or tax advice nor as a recommendation to buy, sell, or hold any particular security. Goldman Sachs believes the data above is accurate, but does not verify its accuracy independently and does not warrant or guarantee that it is accurate or complete. Goldman Sachs has no obligation to provide any updates or changes to the data. No investment decisions should be made using this data.