The Credit Cost Surge Isn’t the Problem — It’s the Symptom

The mortgage industry is once again focused on cost. Recent data from Equifax shows a sharp shift in the economics of credit reporting. Since 2020, the cost of credit scores used in mortgage underwriting has climbed from less than a dollar to roughly $12 per score, now representing a meaningful share of total credit-report expenses.

At the same time, loan conversion has collapsed. What was once a 65% close rate has fallen to nearly 35%.

The response has been predictable: reduce cost, delay credit pulls, and explore alternatives like VantageScore 4.0 following policy changes from the Federal Housing Finance Agency and the GSEs.

But this framing misses the point.

The problem isn’t the cost of credit. The problem is that the industry is paying more for decisions built on data it doesn’t fully trust.

The Changing Economics Of Mortgage Credit

Credit used to be a marginal cost. Today, it’s a controlled expense.

At a 35% close rate, lenders are now absorbing the majority of their upfront costs, including credit, processing, and underwriting on loans that never fund. In many retail environments, total sunk cost per failed file can reach into the thousands when labor, vendor fees, and verification services are included.

Operationally, the pattern is consistent: in many retail pipelines, more than half of originated files require re-verification or data reconciliation after initial qualification before a final decision is reached.

The industry has interpreted this as a pricing problem. In reality, it’s a certainty problem.

Paying More For Less Certainty

Credit scores summarize risk. They don’t validate truth.

They sit atop fragmented borrower data, income from payroll systems, assets from bank statements, and employment from verification vendors, all of which are reviewed at different points in the process.

When those inputs align, decisions are straightforward. Increasingly, they don’t.

Consider a common scenario: a borrower qualifies based on initial documentation and passes an early credit check. Later, during verification, income fails to reconcile across sources. The file stalls or collapses.

By that point, multiple services have already been triggered, including credit.

The system compensates with rework: additional pulls, re-verification, and layered conditions. Cost accumulates, but certainty does not.

The result is a growing disconnect: higher spend with no corresponding improvement in decision reliability.

Fallout Is A Data Failure, Not A Market Cycle

A 35% close rate reflects more than market conditions; it signals structural inefficiency.

Every fallout loan carries a layered cost:

• Credit reports pulled, often more than once
• Verification services reordered as data changes
• Appraisals initiated on unstable files
• Underwriting cycles that never resolve

At scale, the majority of upfront origination spend is now tied to loans that do not close.

These loans are not failing because of credit scores.

They fail because the borrower’s financial profile does not hold up under verification: Income mismatches across sources. Assets fail documentation standards. Data contradicts supporting documents. And these issues are typically discovered late, after cost has already been committed.

Fallout is not a downstream problem. It is the late-stage discovery of upstream data inconsistency.

Why Cost Strategies And Score Competition Miss The Point

Lenders are responding rationally, just not at the root.

Cost strategies are evolving: delaying credit pulls, reducing bureau usage, and tightening when expenses are incurred. At the same time, VantageScore 4.0 is expected to introduce pricing pressure as FHFA policy evolves.

These approaches can reduce marginal cost in specific scenarios. Early filtering, for example, can work when a portion of applicants is clearly non-viable.

But most fallout does not occur at the extremes. It occurs in loans that appear viable early and fail later due to inconsistent data.

In those cases, changing the timing of a credit pull or the scoring model itself does not change the outcome. It shifts when the cost is realized.

Whether using FICO or VantageScore, the output is only as reliable as the inputs. If the underlying data is inconsistent, the decision remains unstable.

The industry has optimized its data scoring. It has not solved the question of how it validates it.

The Shift To Upstream Decisioning (And Its Risk)

The next shift is already underway: moving decisions earlier.

Lenders are using lighter data, one- or two-bureau checks, and preliminary signals — to screen borrowers before incurring the full tri-merge costs. The intent is clear: reduce spend on non-viable files.

But this introduces a critical dependency: data accuracy.

The industry believes it is moving decisions earlier. In many cases, it is moving flawed data earlier. This creates two risks:

• Qualified borrowers are screened out based on incomplete information
• Unqualified borrowers advance further, accumulating costs before failing

The result is not efficiency. It is faster misclassification.

From Scoring Data To Verifying It

The structural gap is not in scoring; it’s in verification timing.

Today, borrower data is validated only after key decisions have been made, and costs have already been incurred. A different model is emerging: the pre-verified borrower profile, a reconciled, decision-ready dataset built before underwriting begins.

In practice, this means aligning income, assets, and employment across sources upfront, using payroll data, bank transactions, and verification tools to resolve inconsistencies before a credit decision is finalized.

Some lenders are beginning to move in this direction, though adoption remains uneven. Integration across systems, vendor coordination, and workflow changes are real constraints.

There are tradeoffs. Upfront verification can introduce additional steps and shift effort earlier in the process. But it fundamentally changes the economics.

When borrower data is reconciled upfront:

• Credit is pulled once, not repeatedly
• Underwriting cycles compress
• Fallout is reduced before cost accumulates

The advantage is not lower unit cost. It is decisions built on data that hold up the first time.

Conclusion

The surge in credit score costs is not the problem. It is a signal.

A signal that the industry is relying on increasingly expensive processes to compensate for a more fundamental gap: the absence of a verified, decision-ready view of the borrower.

Until that gap is addressed, cost-control efforts will remain incremental, and conversion will remain constrained.

Faster decisions won’t define the next phase of mortgage innovation. It will be defined by decisions built on data that holds up the first time.