ONTARIO, Calif.—Record new cars sales won’t continue, and that means loan growth could stall, says one expert who is urging credit unions to hire a data scientist in the near future.
CU Direct, as well as a number of automotive industry analysts, are projecting it’s likely in early 2017 that new cars sales will plateau and begin to decline, moving away from what is expected to be a record 17.7 million new car units sold this year.
“Credit unions need to make sure that as we move forward we can sustain our market share as new car sales come down,” said Michael Cochrum, executive lending advisor at CU Direct, during the company’s latest State of the Auto Lending Market webinar. Credit unions are growing auto loans at a faster pace than all other lenders, and are approaching 21% share of the entire market.
As the market changes next year, Cochrum said CUs must be ready to adjust and to “optimize” their auto lending. He emphasized that using analytics and auto decisioning can not only bring credit unions more loans than a manual process, they can also reduce CU lending costs, allowing credit unions to lower rate and compete even more effectively.
Auto Decisioning
Cochrum pointed to statistics for CU Direct credit unions that illustrate the advantages of auto decisioning and analytics.
“Today credit unions fund about 50% of the applications they approve,” he said. “However, CU Direct credit unions fund 60%-80% of the apps they system approve.”
CU Direct data also show that moving to system-driven loans, compared with manual decisions, generally increases loan growth by three percentage points.
Cochrum addressed some of the “myths” credit unions have regarding manual and automated loan decisions, saying a big mistake is believing underwriters make better decisions than computers—something disproven by CU Direct data, he said.
Michael Cochrum
“The fact is, underwriters are often distracted by unrelated information, where computers are not,” he said.
Cochrum said that through his work with credit unions over several years he has learned that the industry does not do well with analytics.
“The sad fact is that hackers can often get to the information that a credit union has better and faster than we can internally as a business user,” Cochrum said. “We need to be able to construct ways that our business users can access information to make intelligent decisions. But we need to purchase the appropriate tools to do that.”
For those reasons, Cochrum believes the next hire a credit union makes should be a “data scientist.”
Important Hire
“A data scientist is a combination of a mathematician and an IT person. That is going to be one of the most important hires you make in the next 12 to 24 months,” he said.
Other advice from Cochrum to boost auto lending:
- “Build a custom default model. It will improve the accuracy of your decisions and by doing so you will able to plug that information into your LOS to make better decisions on the fly and more accurately assess your risk. It will also help you prepare for CECL (FASB’s current expected credit loss model) calculations that are coming.”
- “Optimize your decision-making. CU Direct Advisory Services have helped our credit unions increase auto decision by 300% on average—taking them from about 10% auto decision to 30%-40% with very little error. In fact, the models we are building for our credit unions today have less than a 1% error rate after the new rules are in place.”