The research behind what we do
Most people trust the first mortgage rate they’re given. Our research shows that’s exactly how they end up paying too much. Here’s the data behind what we do — and how we know it works.
of borrowers overpaid their mortgage in 2025
a year in avoidable costs for the typical borrower (eight-year average)
banks and credit unions in our annual mortgage survey pool
estimated annual overpayment burden since 2022
The Problem
Why most borrowers overpay — and why it’s not their fault
A typical mortgage contract spans 50 pages of technical and legal language. Borrowers must review it within a 48-to-72-hour window before a required rate lock, while simultaneously navigating school districts, inspection reports, and closing timelines. Under that pressure, nearly 90% rely on a rate from their real estate agent’s referral, their current bank, or the first quote they receive.
Those sources offer convenience. They don’t offer competition. Mortgage brokers acting on agent referrals are under no legal obligation under Dodd-Frank to find the lowest available rate. The result, across 3.2 million loan originations analyzed from the 2025 Home Mortgage Disclosure Act dataset, is that 87% of borrowers paid more than the best rate available for their borrower profile — not because competitive pricing didn’t exist, but because the default search process is structured in ways that systematically prevent borrowers from finding it.
The aggregate cost: roughly $11 billion in excess annual payments on 2025 originations alone, or approximately $247 billion over the life of those loans. For the typical borrower, that’s $3,343 a year in avoidable costs — a straight-line average over the first eight years — and $78,186 over the full life of the loan.
The hidden homeownership tax is not a fee disclosed at closing. It’s the gap between what borrowers pay and what the market could deliver if competition were allowed to function as intended — extracted one basis point at a time from monthly payments most households have no way to benchmark.
How We Benchmark
The Bankrate mortgage auction
The benchmark for competitive mortgage pricing comes from Bankrate’s real-time mortgage marketplace — a live exchange where lenders and brokers compete simultaneously for individual borrower business.
How the auction works
When a prospective borrower submits a query, they provide key underwriting inputs required to generate a binding offer: FICO band, debt-to-income ratio, household income, property value, and loan amount. Bankrate anonymizes these inputs and presents them to participating lenders — including banks, credit unions, and non-bank originators — who submit competing bids ranked by estimated eight-year loan cost.
The eight-year cost metric — roughly how long the typical homeowner holds a mortgage before selling or refinancing — standardizes pricing by incorporating the full economic cost of the loan: interest rate, discount points, origination charges, lender fees, and other closing costs, into a single comparable figure. This neutralizes the wide range of pricing structures lenders use and allows like-for-like comparison across offers that allocate costs differently over time.
Binding offers, not estimates
Unlike conventional rate surveys or lead generation businesses, offers submitted through the Bankrate platform are binding — conditional only on the accuracy of the borrower’s inputs and verifying underwriting requirements — and are honored by the winning institution at origination. These bids are directly comparable to the terms at which loans are actually closed in the market.
Across a sample of Bankrate winning bids in 2025, 35.7% of borrowers closed at or better than their Bankrate quote — 15.3% at the exact quoted rate and 20.4% below it. Among those who closed higher, 49% had submitted inaccurate profile information that was corrected at underwriting, and the remainder chose to pay fewer discount points post-auction, accepting a higher rate in exchange for lower upfront costs.
External benchmark: the Bankrate Monitor
To account for variability in auction participation and provide an independent check on pricing quality, Bankrate also surveys mortgage rates through the Bankrate Monitor (BRM) National Index — a weekly survey series incorporated into the Federal Reserve Economic Data (FRED) system. Throughout 2025, we surveyed more than 600 banks and credit unions for mortgage rates, with an average of 300+ institutions sampled each week; not every institution is included in every weekly survey. The broader BRM total of 850+ institutions includes banks and credit unions surveyed across all product categories, including deposits. Against this benchmark, Bankrate’s best offer was compared each day to the institutions surveyed that day: on an average day during 2025, Bankrate’s eight-year cost was lower than about 99.7% of those institutions.
✓ On an average day during 2025, Bankrate’s auction delivered an eight-year cost lower than about 99.7% of the mortgage banks and credit unions surveyed that day, drawn from an annual pool of 600+ institutions.
The Data
Four independent sources behind our overpayment analysis
To measure the scope and cost of mortgage overpayment at scale, this analysis draws on four distinct data sources. Each contributes a different dimension: market pricing, origination outcomes, household wealth context, and credit profile calibration.
- 1
Bankrate mortgage auction
Real-time binding bids from competing lenders, ranked by eight-year loan cost. Serves as the competitive benchmark against which actual origination outcomes are measured.
- 2
HMDA loan-level data
3.2 million mortgage originations from the 2025 Home Mortgage Disclosure Act dataset, administered by the CFPB and the Federal Financial Institutions Examination Council (FFIEC). Provides actual contracted rates, borrower characteristics, and loan terms.
- 3
Federal Reserve SCF
The Survey of Consumer Finances provides household wealth and income context, enabling analysis of how overpayment affects wealth accumulation across the income distribution.
- 4
Fannie Mae & Freddie Mac
Loan-level datasets from both GSEs supply FICO score distributions and underwriting attributes used to calibrate credit profile matching across conventional, FHA, and VA borrowers.
The Methodology
How we measure overpayment
Our overpayment analysis compares the rates borrowers actually received at origination against the competitive offers available to them at the time through Bankrate’s real-time mortgage marketplace.
To ensure an apples-to-apples comparison, each loan in our dataset is matched to Bankrate auction bids for a borrower with a similar credit and financial profile — controlling for credit score, debt-to-income ratio, loan size, loan type, and several other underwriting characteristics. This means we’re not comparing averages to averages. We’re asking: given exactly who this borrower is, what was the best rate they could have gotten?
Rather than a simple yes/no classification, the model assigns each borrower a continuous probability of overpayment, accounting for the fact that market rates vary month to month and that some borrower characteristics must be estimated from related datasets. Costs are measured over an eight-year horizon and include not just interest but fees, points, and other closing costs, so rate-shopping tricks like buying down the rate are properly accounted for.
Sensitivity testing across alternative specifications consistently produces an overpayment rate above 83%, confirming the core finding is not driven by any single methodological assumption.
✓ The full technical methodology is available in our published research: Fellowes & O’Connor, “The Hidden Homeownership Tax,” Bankrate Research, 2026.
What This Means for You
Translating overpayment into real savings
Annual and lifetime savings
The $3,343 in annual savings available through competitive auction pricing is derived from the probability-weighted expected excess cost across the 2025 HMDA cohort — the average gap between what borrowers paid and what Bankrate’s benchmark delivered for an equivalent risk profile, expressed as a straight-line average over the first eight years rather than a single-year out-of-pocket amount. Across that eight-year holding period, it totals approximately $27,000. Projected over each loan’s full scheduled term, and assuming the loan is held to the end of that term, the savings reach $78,186; borrowers who sell or refinance earlier realize less.
Borrowing capacity
The $3,343 in annual savings translates to approximately $47,000 in additional mortgage borrowing capacity at prevailing 2025 rates — enough, in many markets, to move a household from renting to owning or into a meaningfully different home.
Retirement impact
Redirected into a diversified equity portfolio at historical average returns over a 30-year loan term, the same monthly savings of approximately $279 compounds to roughly $341,000 — more than three times the Federal Reserve’s estimate of the median American household’s total retirement savings.
Claims & Sources
Every claim on this page, sourced
We hold every public-facing claim to the same standard we hold our rate data: traceable, specific, and honest about what it does and doesn’t cover.
| Claim | As stated |
|---|---|
| 87%Overpayment rate | 87% of borrowers paid above the competitive market rate. Source: Bankrate Research, “The Hidden Homeownership Tax” (2026). The study's finding is that approximately 87% of 2025 mortgage borrowers overpaid. Based on 2025 purchase and refinance originations for which a comparable Bankrate benchmark was available, not all borrowers. “Overpaid” means the borrower's loan had a higher estimated 8-year cost (including points, fees, and interest) than a loan benchmarked by Bankrate for borrowers with a similar credit profile, and does not reflect an independent or market-wide determination that any individual overpaid. Estimates are subject to methodological uncertainty arising from the imputation of borrower credit attributes. Individual results vary. |
| $3,343Annual avoidable cost | The typical borrower overpays $3,343 a year in excess mortgage costs. Source: Bankrate Research, “The Hidden Homeownership Tax” (2026). Based on 2025 purchase and refinance mortgage originations for which a comparable Bankrate benchmark was available, not all borrowers. The figure compares the estimated total cost of each borrower's actual loan, as reported under HMDA, against the estimated total cost of a loan benchmarked by Bankrate for borrowers with a similar credit profile over an assumed 8-year loan duration, including points, fees, and interest. “Overpay” refers to a higher modeled 8-year total cost on the actual loan, and does not reflect an independent or market-wide determination that any individual overpaid. The $3,343 figure is the average modeled overpayment across all borrowers in this population, counting those who did not pay more as $0, expressed as a straight-line average over the first eight years (the total modeled 8-year overpayment divided by eight), not a single-year out-of-pocket amount. Individual results vary. |
| $65BAnnual overpayment burden | $65 billion lost annually across U.S. mortgages originated since 2022. Source: Bankrate Research, “The Hidden Homeownership Tax” (2026). The $65 billion figure is a modeled estimate. It represents the estimated combined annual excess cost currently paid across mortgages originated from 2022 through 2025, not the cost of any single year's originations. It is a conservative lower bound (earlier loans are excluded) and reflects only those 2022 through 2025 originations for which a comparable Bankrate benchmark was available. For each loan, the estimated 8-year cost (including points, fees, and interest) is compared against Bankrate's most competitive offer for a similar underwriting cohort, and the excess is summed across all such loans and expressed on an annual basis. “Above-market pricing” means a higher cost than that Bankrate benchmark, which is a best-available benchmark, not a typical offer; it does not reflect an independent or market-wide determination. FICO scores are not reported under HMDA and, together with some debt-to-income and loan-to-value data, are estimated using cohort-matched reference data. |
| $11BExcess annual payments | Roughly $11 billion in excess annual payments on 2025 originations alone. Source: HHT study. 2025 HMDA cohort, 3.2 million originations. Period: 2025 loan originations Scope / limitations: Annual payment basis. Lifetime equivalent is ~$247B across the same cohort over 30-year terms. |
| ~9 in 10Originated mortgages beaten | During 2025, Bankrate's auction generated lower rates than ~9 in 10 of all originated mortgages in the U.S. in 2025. Source: Bankrate Research, “The Hidden Homeownership Tax” (2026). The study found that Bankrate's most competitive 2025 offer had a lower estimated 8-year cost than approximately 87% of the comparable mortgages evaluated. “Comparable mortgages” means 2025 purchase and refinance originations reported under HMDA for which a comparable Bankrate benchmark was available (approximately 3.16 million evaluated loans), not all U.S. mortgage originations. For each loan, its estimated 8-year cost (including points, fees, and interest) is compared against Bankrate's most competitive offer for a similar underwriting cohort. That most competitive offer is a best-available benchmark, not a typical offer, and not every borrower would obtain it. FICO scores are not reported under HMDA and, together with some debt-to-income and loan-to-value data, are estimated using cohort-matched reference data. Individual results vary. |
| 99.7%Beat rate | During 2025, Bankrate's purchase & refinance mortgage 8-year costs were lower than 99.7% of 600+ surveyed banks & credit unions. Comparison based on Bankrate's most competitive daily purchase and refinance mortgage offers versus a third-party survey of bank and credit union offers during 2025, comparing similar 30-year fixed-rate loans. Bankrate figures reflect quoted rates users selected on Bankrate's rate tables, not necessarily final or funded loans, using loan-amount and credit-score ranges centered on the survey's parameters to capture sufficient volume for a reliable daily comparison. More than 600 unique banks and credit unions were surveyed over the year (with about 300 surveyed each week), and Bankrate's best offer was compared each day against the institutions surveyed that day. On an average day, Bankrate's 8-year cost (including interest, points, and fees, net of lender credits) was lower than about 99.7% of those institutions. Surveyed institutions represent roughly half of U.S. bank and credit union mortgage originations, not all banks and credit unions. Individual results vary. |
| $78,186Lifetime savings | During 2025, Bankrate's most-competitive offers were projected to save the average U.S. mortgage borrower $78,186 over the full life of their loan. Source: Bankrate Research, “The Hidden Homeownership Tax” (2026). Based on over 3.1 million individual 2025 mortgage originations reported under HMDA that met inclusion criteria and had a comparable Bankrate benchmark available, not all U.S. mortgage borrowers. For each loan, the projected lifetime cost of the borrower's actual originated loan is compared against Bankrate's most-competitive offer within the same underwriting cohort for the period. The most-competitive offer may not have been available throughout the entire period. Lifetime cost includes total interest over the full scheduled loan term plus upfront lender costs, net of credits, and reflects points and fees; principal is excluded. The $78,186 figure is the projected average across all evaluated loans and assumes each loan is held to the end of its full term. Savings would be lower for borrowers who sell or refinance earlier. FICO scores are not reported under HMDA and, together with some debt-to-income and loan-to-value data, are estimated using cohort-matched reference data. Individual results vary. |
| $2,462Avg. refinancer overpayment (annual) | Refinancers overpay $2,462 a year on their mortgage on average. Source: Bankrate Research, “The Hidden Homeownership Tax” (2026). Based on 2025 refinance mortgage originations for which a comparable Bankrate benchmark was available, not all refinance borrowers. The figure compares the estimated total cost of each borrower's actual loan, as reported under HMDA, against the estimated total cost of a loan benchmarked by Bankrate for borrowers with a similar credit profile over an assumed 8-year loan duration and includes points, fees, and interest. “Overpay” refers to a higher modeled 8-year total cost on the actual loan, and does not reflect an independent or market-wide determination that any individual overpaid. The $2,462 figure is the average across all refinance borrowers in this population, counting those who did not pay more as $0, expressed on an annual basis. Individual results vary. |
| $205Avg. refinancer overpayment (monthly) | Refinancing through Bankrate could save the average refinance borrower $205 a month — $2,462 a year — compared to the rate they actually received. Source: Bankrate Research, “The Hidden Homeownership Tax” (2026). Based on 2025 refinance mortgage originations for which a comparable Bankrate benchmark was available, not all refinance borrowers. The figure compares the estimated total cost of each borrower's actual loan, as reported under HMDA, against the estimated total cost of a loan benchmarked by Bankrate for borrowers with a similar credit profile over an assumed 8-year loan duration and includes points, fees, and interest. “Overpay” refers to a higher modeled 8-year total cost on the actual loan, and does not reflect an independent or market-wide determination that any individual overpaid. The $2,462 figure is the average across all refinance borrowers in this population, counting those who did not pay more as $0, expressed on an annual basis. Individual results vary. |
| $700M+Member savings per year | $700M+ saved by Bankrate members every year. $700M+ savings reflects total projected savings across Bankrate users on deposit and mortgage products. The average rate on each funded product shown on Bankrate is compared against a national benchmark: Freddie Mac 30- and 15-year averages for mortgages; FDIC published rates for deposits, with mortgage savings projected over an assumed 77-month hold and deposit savings over an assumed 6-year account life. These are combined Bankrate user projected savings over the products' estimated periods, not amounts realized within a single calendar year. The $700M+ figure is the average annual total from 2020 through 2025. |
| 600+Institutions surveyed | 600+ total home-lending banks and credit unions surveyed annually (2025, purchase + refinance combined). Source: Bankrate Monitor (BRM) National Index Period: 2025, ongoing Scope / limitations: Annual pool. Weekly comparison basis is 300+. Total BRM including deposit-only is 850+. |
| $47KBorrowing capacity | $3,343 in annual savings translates to approximately $47,000 in additional mortgage borrowing capacity. Source: Derived. $3,343/yr annualized at prevailing 2025 30-yr fixed rates. Period: 2025 Scope / limitations: Illustrative calculation. Actual capacity depends on lender, credit profile, DTI, and prevailing rates at time of application. |
| $341KRetirement impact | Redirected into a diversified equity portfolio, the monthly savings compounds to ~$341,000 over 30 years. Source: Derived. ~$279/mo invested at historical average equity returns over 30-year term. Period: Illustrative / 30-yr projection Scope / limitations: Illustrative only. Past investment returns do not guarantee future results. Individual outcomes vary based on asset allocation, timing, and fees. |
Claims sourced from the HHT study reflect findings as of the study publication date. Past performance is not indicative of future results. Bankrate is not a lender.
Public DataBankrate rate data is available through the Federal Reserve’s FRED platform
While institution-level rate data is available exclusively to Bankrate members, we share national averages with the Federal Reserve. The Bankrate Monitor (BRM) National Index and Bankrate Consumer Polls are hosted by the Federal Reserve Economic Data (FRED®) platform, maintained by the Federal Reserve Bank of St. Louis. Researchers and policymakers can retrieve, graph, and download these series alongside other leading economic indicators.