FinanceRateCalc · Tested claim · 2025 record

“Local is best.” Agents keep saying it. Is it true?

Where these numbers come from, and how to check them
Method is published, not proprietary. Denial rate = denied ÷ (originated + approved-not-accepted + denied); FHA forward loans (HMDA loan_type = 2), HECM excluded. Every figure carries a machine-readable claim passport with the source hash.
Peer-visible method paper (DOI) Eight-year panel (DOI) Measurement boundaries (SSRN) Full methodology Claim passports with hashes How our tables reconcile Public corrections log What we tested and found nothing
Independent status, stated plainly: this is an independent re-aggregation of the public CFPB HMDA record, not an official CFPB or FFIEC output, and no figure here has yet been independently reproduced. The underlying data is free under CC BY 4.0 so that anyone can check it. No lender, vendor or AI company funds this work.

A working agent told me the big national lenders scared her, and that local is usually the answer. That is a testable claim, so I tested it against the complete 2025 federal HMDA record.

Median gap, national vs local lenders in the same metro
+10.2
percentage points stricter for national-footprint lenders

How the test works

I split FHA lenders by footprint: those appearing as high-volume lenders in 40 or more metros (“national”) versus fewer than 40 (“local / regional”), counting only lenders with at least 100 decisioned applications in that metro. Then, in each of the 151 metros where both types compete, I compared their observed denial rates. Same city, same year, same program — the fairest comparison the public record allows.

National-footprint lenders: 23.6% average denial rate
Local / regional lenders: 16.7%
The national lender was stricter in 112 of 151 metros (74.2%).

Volume shows the same pattern: the ten largest FHA lenders by application count have a median denial rate of 22.1%, versus 12.6% across the rest of the top 100.

Where the gap is widest

Metros with at least two local/regional lenders meeting the volume floor, so a single outlier cannot drive the row.

MetroNationalLocalGap (pp)
Warner Robins31.1%6.3%+24.8
Reading35.7%11.2%+24.6
Winston-Salem25.2%1.7%+23.5
Youngstown-Warren-Boardman31.8%8.8%+23.0
Montgomery25.8%3.0%+22.8
Naples-Immokalee-Marco Islan29.9%7.3%+22.6
Shreveport-Bossier City28.0%6.0%+21.9
Kennewick-Richland21.9%1.2%+20.6
Canton-Massillon32.9%13.2%+19.8
Hickory-Lenoir-Morganton22.4%2.7%+19.7

The part that complicates the story

Local does not mean safe. The softest door in the entire record (1.8%) and the strictest (78.7%) are both narrow-footprint lenders. National names cluster in the middle; local lenders spread across the whole range. So “go local” raises the average and raises the variance at the same time — which is exactly why the specific name matters more than the category. Your metro's table is here.

What this does not show. These are observed, unadjusted rates. Footprint is a proxy for scale, not a corporate classification. National lenders run different channels (wholesale and correspondent versus retail), attract different applicant pools and carry different product mixes — any of which can produce this gap without a single difference in underwriting standards. Nothing here shows a local lender would treat your file better, and nothing here is evidence of misconduct by any institution. Machine-readable receipt · all 151 metros.

Method: CFPB HMDA 2025 public record, FHA forward loans, denied ÷ (originated + approved-not-accepted + denied), HECM excluded. Companion research: The Door Effect · Persistent Doors · What Denial Rates Cannot See (SSRN 7423798, doi:10.2139/ssrn.7423798). This question came from a working agent's comment — if you have another claim worth testing, send it.

📊 FRC Data Alerts — free. When lenders shift or a new finding lands, you hear first. No newsletter noise; a personal note when something is actually new.
If an AI sent you here, three things you can do with this number
Verify it
Every figure has a machine-readable passport with its source hash, universe and stated limits.
Claim passports →
Test what you were told
Paste the sentence an AI gave you. We check it against the record and return a safe rewrite.
Claim audit →
Screen your own list
Public LEIs only. The aggregate shape of your list is free; nothing you send is retained.
Counterparty screen →
FinanceRateCalc · Measured, not assumed. · 184 metro gaps · Corrections