The Two-Track Future of Credit Scoring: Open-Loop vs. Closed-Loop Models
Open-loop credit scores like FICO are no longer the only yardstick in consumer lending. Closed-loop scores, such as Cash App’s “Cash App…
The Two-Track Future of Credit Scoring: Open-Loop vs. Closed-Loop Models

investopedia
Open-loop credit scores like FICO are no longer the only yardstick in consumer lending. Closed-loop scores, such as Cash App’s “Cash App Score,” now operate inside specific platforms and shape credit decisions for tens of millions of users. The result is a split system: open scores still anchor mortgages and capital markets, while platform-specific scores decide how much credit a customer gets inside a given app. That shift benefits some borrowers, frustrates others, and forces lenders, bureaus, and fintechs to make new tradeoffs.
1. How FICO’s role began to change
In 2018, Congress passed the Economic Growth, Regulatory Relief, and Consumer Protection Act. Among other things, it required the FHFA to open a path for credit models beyond FICO for loans sold to Fannie Mae and Freddie Mac. That ended the long-held assumption that one score would remain mandatory for mortgages, even though rollout would be slow.
By mid-2025, the FHFA moved from planning to execution. Under Director Bill Pulte, it allowed lenders to use VantageScore 4.0 alongside existing FICO models for Fannie and Freddie loans, while keeping tri-merge credit files in place. The FHFA and VantageScore argue that this will expand mortgage access for younger and thin-file borrowers without weakening risk controls.
Over the same period, FICO’s business changed. Its share price rose from under $200 in 2018 to above $2,000 by 2024 as investors priced in its central role in credit decisioning. Mortgage score pricing also climbed, from well under $1 per pull historically to several dollars per score. In parallel, FICO pushed toward a direct licensing model, with mortgage scores priced around $10 in some channels.
2. Open loop and closed loop scores
Open-loop scores are built to move across lenders and products.
They typically:
- Rely on bureau data from many furnishers and tradeline types.
- Let lenders compare risk, price loans, and sell them into secondary markets using a shared standard.
- Appear in bank apps and third-party tools, giving consumers a way to compare offers and plan large purchases.
FICO and VantageScore both fit this model. FICO uses the traditional three-bureau framework. VantageScore 4.0 adds trended data and some non-traditional inputs, such as rent and telecom histories, and claims to score tens of millions more people than older models.
Closed-loop scores serve a different purpose. They are built for one platform, not the whole market.
They usually:
- Rely on first-party data from a single app, wallet, marketplace, or issuer.
- Answer “how much credit should we extend here” rather than “how risky is this borrower everywhere.”
- Update frequently using cash flow and behavior that may never reach the bureaus.
Cash App Score, internal BNPL scores, and large-bank proprietary models all fall into this group. Cash App is notable because it now shows its internal score directly to users.
3. Why FICO’s advantage narrowed
Several trends chipped away at FICO’s open-loop dominance:
- Large banks built in-house models that combine bureau data, third-party scores, and their own account histories to manage risk and returns by segment.
- Fintech and BNPL lenders scaled quickly without consistently furnishing data to bureaus, leaving bureau files thinner for younger users and heavy BNPL customers.
- Consumer “credit builder” products sometimes boosted bureau scores without mirroring traditional installment risk, adding noise for certain segments.
The gap widened between what a bureau score shows and what lenders see in their own portfolios, especially for subprime and credit-invisible users. At the same time, many lenders stuck with older FICO versions outside mortgages, reducing the payoff from paying for upgrades.
4. Competition and pricing
Once the FHFA allowed model choice, pricing and distribution shifted.
FICO:
- Moved mortgage scores toward direct licensing, with headline prices near $10 in some channels.
- Unwound parts of its prior bureau distribution structure, changing economics for resellers and large lenders.
Equifax and VantageScore:
- Set VantageScore 4.0 pricing around $4.50 per mortgage score for at least two years, well below FICO’s new list prices.
- Offered VantageScore 4.0 at no charge through 2026 for certain Equifax customers who also buy FICO scores, aiming to drive adoption.
The FHFA frames this as a way to lower borrower costs while keeping familiar tri-merge processes in place.
5. Cash App Borrow and Cash App Score
Cash App now serves more than 50 million users and bundles payments, debit, investing, and short-term credit. Its lending product, Cash App Borrow, has grown fast:
- About $9 billion in Borrow originations in 2024, initially through a Utah partner bank.
- FDIC approval in early 2025 for Square Financial Services to originate and service these loans directly.
- Loss rates under 3 percent, with most revenue from borrowers who repay on time.
Borrow loans are small and short term. Repayment is often automated from incoming Cash App funds, which limits exposure and reduces reliance on bureau scores.
Block has said its underwriting leans on near-real-time cash flow and behavior rather than backward-looking bureau data. Cash App Score draws from:
- In-app transactions such as P2P payments, card spend, ACH deposits, and bill pay.
- Repayment history on Cash App Borrow and Afterpay loans, which are not fully reported to bureaus.
- Other first-party signals like savings and investing activity.
In 2025, Block began testing a feature that lets users see their Cash App Score, view key drivers, and get guidance on raising it to qualify for higher Borrow limits. The score updates often and reflects standing inside the Cash App system, not the wider credit market.
6. Why platforms use closed loop scores
For platforms like Cash App, closed-loop scores offer clear benefits:
- Tighter segmentation based on how users behave inside the app, not elsewhere.
- Faster feedback, allowing limits and pricing to change without waiting on bureau updates.
- Stronger engagement by tying visible scores to product access and in-app actions.
That logic explains why banks and fintechs already rely on internal models, and why user-facing platform scores may spread to other apps.
7. Limits for consumers
Closed-loop scores come with downsides:
- Portability: A strong Cash App Score can unlock cheaper credit in-app but may not help with lenders that rely on bureau scores, since Cash App does not fully furnish repayment data today.
- Scope: Scores trained on short-term advances and BNPL may not map cleanly to longer-term risks like auto loans or mortgages.
- Clarity: Users may struggle to connect a platform score to their broader credit profile, and data-sharing paths are still limited.
Block has discussed giving users more control over their data and making it useful beyond Cash App. Doing so would require clear data-sharing rules, compliance with credit reporting law, and proof that these scores add value outside small-dollar lending.
8. What comes next
Credit scoring is likely to keep running on two tracks:
- Open-loop scores will remain central for mortgages and markets that demand consistency across lenders.
- Closed-loop scores will keep shaping credit inside apps, using cash flow and behavior that never hits the bureaus.
For borrowers, the distinction matters. A score shown inside an app increasingly answers, “how do we see you here.” Bureau scores still decide access to large, transferable products where comparability matters most.
메타데이터
- post_id
- 987c7b4aa712
- slug
- the-two-track-future-of-credit-scoring-open-loop-vs-closed-loop-models-987c7b4aa712
- url
- https://medium.com/@danielychang/the-two-track-future-of-credit-scoring-open-loop-vs-closed-loop-models-987c7b4aa712
- canonical_url
- https://medium.com/@danielychang/the-two-track-future-of-credit-scoring-open-loop-vs-closed-loop-models-987c7b4aa712
- author_url
- https://medium.com/@danielychang
- status
- ok
- fetched_at
- 2026-06-16 19:09:56