Bankrate analyzed 3.2 million mortgage originations in federal housing data against offers available in its marketplace and found that 87% of borrowers likely overpay. Across mortgages originated since 2022, that amounts to roughly $65 billion a year, about $3,343 per household, or $78,186 over the life of a 30-year loan.
That last figure is worth holding next to another one: it exceeds the total retirement savings of the median American household.
But the aggregate is less interesting than the distribution, because the distribution is backwards from what you would expect.
The strongest borrowers do the worst
Among borrowers in the lowest debt-to-income quartile, 91% overpay. In the second-lowest quartile, 92% overpay, the highest rate in the entire dataset. Conventional mortgage holders overpay 89% of the time, carrying lifetime excess costs equal to 23% of their loan balance.
The people with the cleanest financial profiles, the ones with the most negotiating leverage, are the most likely to leave money on the table.
The explanation lies in a distinction that gets collapsed constantly. Your credit profile determines the rate you are eligible for. It does not determine the rate you are offered. Those are separate numbers, and the space between them is the lender's margin.
A strong borrower receives a quote and reads it as a verdict on their qualifications. It is not. It is a pricing decision, made by a lender who knows the great majority of borrowers accept the first offer. Excellent credit guarantees you will be approved. It guarantees nothing about the price, because price reflects competitive pressure, and a borrower who does not shop applies none.
There is a behavioral layer too. For borrowers with weaker profiles, the mortgage process feels like a search for approval, so they may apply in several places out of anxiety and stumble into comparison shopping. A borrower with an 800 score and low DTI never experiences that anxiety, applies once, gets approved, and never learns what else was available.
What the spread actually looks like
The gap being left is not theoretical. LendingTree data covering October 2025 through April 2026 found the average lowest offered rate was 5.93% and the average highest was 6.72%, a 0.79 point spread that translates to $174 a month, or $62,572 over a 30-year loan on the average requested amount.
That is the same borrower, same profile, same period, receiving offers nearly eight tenths of a point apart. And it widens with effort: borrowers who gathered six or more offers saw a 0.98 point spread, meaning the more you look, the further apart the extremes turn out to be. In high-cost states the dollar figures escalate, with average potential savings near $89,000 in Hawaii and around $82,000 in New Jersey and California.
The accordion effect
Freddie Mac's research on rate dispersion, the range of rates offered to similar borrowers on the same day, explains why this got worse rather than better.
Between 2010 and 2021, average dispersion was under 20 basis points, and applying with two lenders saved an average of 10 basis points. When rates rose at their fastest pace in four decades, dispersion more than doubled to about 50 basis points, and the average savings from a second application doubled with it.
Bankrate's Greg McBride called it an accordion effect: as rates rise, the disparity between quoted terms expands.
The implication is uncomfortable. Shopping matters most precisely when borrowers are least inclined to do it. At 6.5%, the natural reaction is that rates are high everywhere and nothing can be done, which is exactly when the spread between lenders is widest and the payoff for looking is largest. The despair and the opportunity peak together.
The excuse that does not survive the data
The standard defense is time pressure. Purchase closings run on deadlines, sellers want certainty, and a buyer under contract has limited bandwidth to solicit quotes.
That explanation collapses against refinance borrowers, who face no deadline, no seller, no competing offers, and unlimited time to shop. They overpaid 79% of the time, against more than 90% of purchase borrowers.
Better, but four in five is not a story about time. It is a story about not knowing the spread exists, or assuming the first quote is the market rate rather than one lender's price.
The myth that stops people
The most common reason borrowers give for not gathering multiple quotes is fear that several credit inquiries will damage their score.
The scoring models were built with this in mind. Multiple mortgage inquiries within a defined shopping window, generally 14 to 45 days depending on the model, are treated as a single inquiry for scoring purposes, precisely so consumers can comparison shop without penalty. The rate shopping is protected. The belief that it is not costs borrowers tens of thousands of dollars.
What actually works
The mechanics are simple enough to fit in a short list.
Get at least three quotes, and preferably five, since Freddie Mac's dispersion data suggests that is roughly where you gain confidence you have found the bottom of the range. Do it inside a two-week window, which keeps every inquiry comfortably inside the shopping period under any scoring model.
Compare Loan Estimates rather than advertised rates. The standardized Loan Estimate form exists so offers can be compared line by line, and the rate alone is incomplete without origination fees, points, and lender credits, which are where an attractive headline rate is frequently recovered.
Include different lender types. Credit unions, independent mortgage banks, and large banks price differently, and a spread of nearly a point across the market means the outlier that helps you may not be the institution you already bank with.
And use the quotes. A written competing offer is the only leverage that reliably moves pricing, because it converts an abstract possibility that you might shop into a documented fact that you did. A mortgage is likely the largest purchase of your life, and it is the one where people do the least comparison shopping. The 87% figure is not evidence that lenders are behaving improperly. It is evidence that a market where most participants accept the first price will produce a wide range of prices, and that the cost of being in the majority is measured in tens of thousands of dollars.