Carrington Labs builds cash flow underwriting models as well as tools for loan and limit sizing, limit management, and early-warning systems for banks and non-bank lenders.
Carrington Labs finds behavioral changes predict loan risk, with new model giving them record 36% weight—highlighting value of spending and income shifts.
This finding underlines the value of behavioral data in assessing a loan applicant’s risk levels.— Jamie Twiss, CEO of Carrington Labs
SYDNEY, NSW, AUSTRALIA, April 17, 2025 /
EINPresswire.com/ -- Research released today by
Carrington Labs, a provider of tailored cash flow underwriting models and credit risk analytics tools for consumer and SMB lenders, showed the predictive value of identifying recent behavioral changes in loan applicants. The weighting of behavioral factors in a recent risk model increased to 36% of total factor weights, a new record for a Carrington Labs model.
In the process of retraining a cash flow underwriting model on a client’s data, the Carrington Labs team found that a series of recently created attributes focused on specific elements of behavioral change were highly predictive, with several placing in the top 30 attributes. The client gave Carrington Labs permission to publish a summary of the findings.
“While we’ve always looked at a range of behavioral factors, this latest generation of cash flow underwriting models tests a wider range of attributes than ever before, and we were surprised to see how many behavioral elements ended up in this particular model,” said Jamie Twiss, the CEO of Carrington Labs. “This finding underlines the value of behavioral data in assessing a loan applicant’s risk levels.”
While Twiss did not disclose the exact behavioral attributes, he did cite two specific categories. The first category includes attributes related to the compressability of discretionary spend. “If an end customer is facing upcoming expenses or a reduction in income, do they reduce their spend in non-essential categories to compensate–and if so, to what extent, and how far in advance do they start? This could be a consumer deferring entertainment expenses, or a small business delaying equipment purchases until the following month,” said Twiss.
The second category included behavioral attributes linked to efforts to clean up a loan applicant’s financials shortly before applying for credit. “If someone is suddenly repaying older debts, or artificially boosting income for a short period of time, it’s a strong indicator of higher risk,” said Twiss.
Learn more about how the Carrington Labs platform works.
About Carrington Labs
Carrington Labs builds cash flow underwriting models as well as tools for loan and limit sizing, limit management, and early-warning systems for banks and non-bank lenders.
It uses contemporary data-science techniques, machine learning, explainable AI, and alternative sources of data to help lenders modernize their decision-making processes, provide the right offers to customers, increase approval rates, and improve margins.
Working across the consumer and small-business lending space, Carrington Labs can pilot a tailored risk model for a lender in days, and onboard a lender in weeks, driving significant improvements in growth and returns.
For more information,
visit www.carringtonlabs.comMatt McLoughlin
Gregory FCA on behalf of Carrington Labs
media@carringtonlabs.com
Visit us on social media:
LinkedIn