Jennifer Henry of Equifax sits down with Andrew Davidson, president of Andrew Davidson & Co. and a leading voice in mortgage analytics, to unpack one of the most misunderstood elements of housing finance: credit scores. They explore what a credit score actually measures, why different models and bureaus produce different results, how VantageScore’s adoption could reshape risk evaluation, and what investors, lenders, and consumers need to know as the industry shifts toward new data sources and scoring frameworks.

What is a credit score and what does it measure?

A credit score is a model applied to a specific credit file to predict the likelihood that a borrower will become delinquent. It is based only on the data included in that credit file, not the consumer’s entire financial life.

Why do credit scores differ between bureaus or scoring companies?

Scores vary because:

• Each bureau holds different underlying data.
• Scoring companies group data differently based on their models.
• The same borrower may fall into different “risk buckets” depending on how the model evaluates attributes (e.g., payment history, utilization, depth of file).

Different models may both predict risk effectively yet categorize borrowers differently.

How does adopting multiple scores (e.g., VantageScore + FICO) affect the industry?

Having multiple accepted scores encourages deeper analysis of:

• How risk is grouped and measured
• Which score is most predictive for different loan types
• How investors calibrate pricing and performance expectations

This shift pushes the industry to understand why scores differ, not just rely on a single number.

How could alternatives to tri-merge (bi-merge or single file) impact lending decisions?

Using fewer files may lower cost and streamline operations, but may reduce visibility into borrower behavior—especially for thin-file or non-traditional applicants. More data generally improves risk grouping.

How does alternative data (e.g., utilities, telco, rental history) influence credit scoring?

Alternative data helps:

• Create a more complete financial picture
• Surface strong repayment behavior not shown on traditional trade lines
• Improve risk assessments for people with non-traditional income patterns or limited credit history

However, adding new data is not enough. Lenders and investors must also understand how that data influences models.

Where can listeners learn more?

Andrew Davidson & Co: ad-co.com
Financial Lifecycle Education (FiCycle): ficycle.org