One mortgage benchmark crossed 7%. Another did not. Both can be right.

For a household borrowing $400,000, the gap between a 6.95% and 7.12% mortgage works out to about $46 a month in principal and interest. Yet those two rates, reported at nearly the same time, do not describe the same slice of the mortgage market.

The Mortgage Bankers Association put its average contract rate for conforming 30-year fixed mortgages at 7.12% for the week ending September 18. Freddie Mac reported 6.95% as of September 17. Then Mortgage News Daily recorded 7.26% on September 23.

That is not evidence that one measure failed. It is a reminder that “the mortgage rate” is not a single national price.

Three windows onto the market

The MBA figure comes from applications submitted through retail and direct-to-consumer channels. Its 7.12% observation applied to conforming loans at an 80% loan-to-value ratio and included an average of 0.73 points, including the origination fee. Its comparable rate was 6.85% for the week ending September 4, so the series rose 0.27 percentage point over two reported weeks.

Freddie Mac’s weekly measure covers qualifying conventional, conforming purchase applications submitted through its underwriting system. Its published 30-year average climbed from 6.71% on September 3 to 6.76% on September 10 and 6.95% on September 17, according to the survey archive. The September 17 release averaged offers made from the preceding Thursday through Wednesday.

Freddie Mac no longer publishes points and fees because those fields are not always supplied by lenders. That makes its headline rate unsuitable for a direct price comparison with an MBA rate carrying a stated points assumption.

Mortgage News Daily operates on a still shorter clock. Its index follows lender rate sheets each business day and uses a standardized, strong-credit scenario. It also adjusts for the cost of points through a proprietary method. Its 30-year reading moved from 6.89% on September 8 to 7.19% on September 17 and 7.26% on September 23.

In short, the figures differ in timing, eligible loans, borrower assumptions and treatment of upfront costs. A weekly average can remain below 7% even as a later daily gauge moves above it.

A rate can hide an upfront price

Points exchange money at closing for a different interest rate. The Consumer Financial Protection Bureau explains that one point equals 1% of the loan amount; paying discount points generally lowers the rate, while lender credits generally do the reverse.

On a $400,000 loan, 0.73 point is arithmetically $2,920. But the MBA’s reported figure includes an origination fee, and it is an average rather than a quote available to every applicant. Comparing headline rates while ignoring points can therefore make the lower-looking loan appear cheaper without establishing that it is.

What the difference does to a payment

For the same hypothetical $400,000, 30-year fixed loan, the standard level-payment calculation is:

payment = P × r × (1+r)^360 / ((1+r)^360 − 1)

Here, P is $400,000 and r is the annual rate divided by 12. Anyone can reproduce the calculation in a spreadsheet by substituting the three rates:

These amounts exclude taxes, homeowners insurance, mortgage insurance, fees and points. The 6.26% case is deliberately hypothetical: Freddie Mac’s observed 6.26% figure on September 17 was its 15-year average, not a 30-year offer. Using it with a 30-year term merely shows the scale of an 0.86-percentage-point difference—about $228 a month on this example loan.

The Fed influences the chain; it does not post the quote

Mortgage rates are priced through financial markets. Treasury yields provide a broad reference, while yields on mortgage-backed securities reflect the returns investors demand for pools of home loans. Lenders then add costs and borrower-specific adjustments before producing a quote.

Federal Reserve policy can influence those market yields by changing short-term rates and expectations about inflation and economic conditions. It does not prescribe the 30-year rate offered to an individual borrower. The Fed’s July monetary policy report, for example, documented elevated inflation and differing measures of price pressure—conditions markets may assess when pricing longer-term debt. That context does not, by itself, prove what caused any particular day’s mortgage move.

The practical reading of a “7% mortgage” headline is therefore narrower than it sounds: it describes one benchmark, for a defined group of applications, during a defined period and under particular cost assumptions. The useful comparison is not one headline against another, but loan offers matched for term, loan type, points, credits and closing costs.