Industry: Consumer Lending Capabilities: Credit Risk Optimization
Unlocking Hidden Credit Capacity: Using Alternate Data for Smarter Credit Decisions
Turning Overlooked Bureau Data and a Lien Clause Into $5.8M in Annual Profit — Without Loosening Credit Standards
Client Profile
Our client is a fast-growing fintech lender specializing in point-of-sale loans for home improvement, with roughly $1.8B in loans under management and monthly originations near $100M. Approve/decline decisions relied on a blend of off-the-shelf bureau scores and a custom score built for the client’s specific borrower population, while line assignment was driven by debt-to-income limits, monthly payment caps, and creditworthiness thresholds. A distinguishing feature of the loan structure was the client’s contractual right to place a lien on the borrower’s property — a powerful, underused backstop against unpaid balances in the event of a home sale, refinance, bankruptcy, or foreclosure.
The Challenge
Despite a sound underwriting framework, the client’s income-driven limit assignment was systematically under-serving a valuable segment: older, low-risk homeowners with substantial equity and likely high net worth.
Consider one representative case: a retired homeowner with a perfect FICO score, no outstanding debt, and full ownership of an $800K property was approved for just $8.5K against a requested $12K. The shortfall traced back to an incomplete income picture — her application likely didn’t capture private pension income or Social Security, both common blind spots in standard income verification.
This wasn’t an isolated case. It pointed to a structural gap: the client’s policy was leaving low-risk, high-equity customers underfunded simply because reported income didn’t reflect their true financial capacity.
The Solution
We saw an opportunity to use the lien provision — already embedded in the loan contract — as the foundation for a smarter, risk-differentiated line extension policy.
Our approach began with a deep dive into individual bureau attributes beyond the scores already in use, testing which factors most effectively differentiated credit risk. We examined debt serviceability by debt type, mortgage recency, and delinquency patterns. The standout finding: revolving/credit-card debt-to-income ratio was the single most powerful differentiator of expected credit loss — and it held true consistently across every FICO segment, not just in aggregate.
Building on this insight, and pairing it with conservative, lien-backed recovery rate assumptions, we designed a tiered line extension policy. Customers were segmented into five risk tiers, with tier 5 (highest risk) subject to the strictest qualifying thresholds across credit card debt-to-income, income, and home equity coverage, and progressively more flexible criteria applied down through tier 4 and below. This let the client extend credit responsibly to genuinely low-risk borrowers — like the retired homeowner above — without loosening standards for higher-risk segments.
As a natural extension of the same analysis, we also identified a narrow cohort of borrowers carrying high credit card debt whose expected losses ran roughly 5x the portfolio average — a clear signal to tighten, not expand, exposure for this group.
Impact
Back-testing validated both recommendations:
Line extension policy: ~2% increase in approvals, delivering an estimated $4M in annual profit benefit
High-risk suppression: Targeted tightening for the highest-loss cohort added a further $1.8M in annual profit benefit
Implications:
This engagement illustrates a recurring theme in credit risk analytics: the data needed to unlock better decisions is often already sitting inside the bureau file or the loan contract — it simply hasn’t been asked the right question yet. A single well-chosen attribute, tested rigorously across risk segments, can outperform broad policy loosening and deliver measurable, low-risk profit gains.