Where Are Mortgage Rates Headed in 2026?

Mortgage rates are the most consequential number in a housing decision and the one most often misread. The 30-year fixed averaged 6.58% in the week ending 23 July 2026 in Freddie Mac’s Primary Mortgage Market Survey, published by the Federal Reserve Bank of St. Louis as series MORTGAGE30US. On a $360,000 loan that is $2,294 a month in principal and interest, before a dollar of property tax, insurance or maintenance. Move the same loan half a point and the payment changes by roughly $120 a month and by about $43,000 across thirty years.

About the rate figures in this guide. The mortgage rates used below are illustrative examples, published on 16 May 2026. Rates move every week. Freddie Mac’s Primary Mortgage Market Survey put the 30-year fixed-rate average at 6.66% and the 15-year at 6.04% for the week ending 30 July 2026, up from 6.58% the week before. Check the current PMMS before relying on any figure here. Last reviewed 31 July 2026.

This article sets out where rates actually stand, the machinery that moves them, which data releases are worth watching, and how a fixed rate compares with an adjustable one at current levels. It does not contain a prediction, for reasons the section on forecasting makes clear.

30-Year Fixed Mortgage RateMonthly average of Freddie Mac’s weekly survey, Jan 2019 – Jul 2026 Source: Freddie Mac Primary Mortgage Market Survey via FRED. Chart: The Housing Signal.30-Year Fixed Mortgage RateMonthly average of Freddie Mac’s weekly survey, Jan 2019 – Jul 20266.58%latest weekly reading, week ending July 23, 2026 2% 3% 4% 5% 6% 7% 8% 20192020202120222023202420252026Source: Freddie Mac Primary Mortgage Market Survey via FRED. Chart: The Housing Signal.

Where Rates Stand in Mid-2026

The monthly average of the 30-year fixed has spent the past year in a narrow band between 6.05% and 6.59%. For context, the same series bottomed at 2.68% in December 2020 and peaked at 7.62% in October 2023. Today’s level sits closer to the middle of that seven-year range than to either end, which is worth holding on to when a headline describes rates as either historically high or finally falling.

The shorter-term path matters more for anyone transacting this year. Rates eased through the second half of 2025 and reached 6.05% in February 2026, the lowest monthly average since the 2022 repricing. They have risen in every month since: 6.18% in March, 6.33% in April, 6.44% in May, 6.49% in June and 6.51% in July, with the latest weekly reading at 6.58%. That is a drift of roughly half a point off the February low, not a break in either direction, but it is a drift in one direction and it has been consistent.

The 15-year fixed averaged 5.96% over the same week, a spread of about 0.62 points below the 30-year. That gap is the compensation lenders require for the longer commitment, and it is the single largest lever available to a borrower who can carry the higher payment.

Why the Fed Does Not Set Your Mortgage Rate

The most persistent misunderstanding in this subject is that the Federal Reserve sets mortgage rates. It does not. The Fed sets the federal funds rate, an overnight rate between banks. Your mortgage is a thirty-year obligation, and its price is set in the bond market by investors deciding what yield they need to hold thirty years of exposure.

The chain runs through the 10-year Treasury. Mortgage-backed securities compete for the same investor capital as Treasuries, so the 30-year fixed tracks the 10-year yield closely, plus a spread that covers prepayment risk, credit risk and servicing. The 10-year yield in turn reflects what investors expect inflation and growth to do over the coming decade. Fed policy influences those expectations, which is the real transmission mechanism, but it is indirect and it operates with a lag.

The practical consequence is that mortgage rates frequently move before an FOMC meeting and barely move on the day of the decision itself, because the decision was already priced in. They can also move against the Fed. A rate cut delivered into an environment where investors think inflation is reaccelerating can be followed by mortgage rates rising, because the long end of the curve is pricing the inflation, not the policy rate. Anyone waiting for a cut in order to buy should understand that the cut is not the event that moves their quote.

The spread between the 10-year Treasury and the 30-year fixed is itself variable. It widened sharply through 2022 and 2023 as prepayment uncertainty rose and the Fed stopped adding mortgage-backed securities to its balance sheet, and it has narrowed unevenly since. A borrower can therefore see mortgage rates fall while Treasury yields are flat, purely because that spread compressed.

The Inflation Data That Actually Moves Rates

Two inflation measures dominate. The Consumer Price Index, published monthly by the Bureau of Labor Statistics, is the one that reaches the headlines and the one bond markets react to fastest. The Personal Consumption Expenditures price index, and specifically core PCE, which strips out food and energy, is the measure the Federal Reserve names in its own framework and the one its 2% target refers to.

The two do not print the same number. CPI uses a fixed basket and weights shelter heavily, while PCE adjusts for the way households substitute between goods as prices change and draws its weights from a broader set of business data. CPI therefore tends to run somewhat hotter than PCE, and the gap widens when housing costs are moving faster than everything else. A month in which CPI surprises to the upside and core PCE does not is a month in which the initial bond-market reaction often partly reverses.

For a mortgage borrower the useful discipline is to watch the direction of core PCE over three to six months rather than any single monthly print. One hot month is noise. A run of three moving the same way is the kind of thing that reprices the long end of the curve, and it is the long end that sets your rate.

The monthly employment report matters for the same reason and in the opposite direction. Strong payroll growth and rising wages suggest an economy that does not need support and can sustain price pressure, which pushes yields up. Weak employment data pulls them down, because investors begin pricing policy easing and slower growth. Nothing here is a rule; it is a tendency, and it is regularly overwhelmed by whatever else is happening in global markets on the day.

Fixed Against Adjustable in This Market

When fixed rates sit well above their recent lows, adjustable-rate mortgages become more visible in lender marketing. The argument is straightforward: a lower introductory rate for the first five years, with the expectation of refinancing or selling before the first reset. The arithmetic underneath that argument deserves examination before the decision, not after.

Fixed against a 5/1 ARM on a $360,000 loan: what happens at the first reset
Scenario Rate Monthly principal & interest Against the fixed payment
30-year fixed, entire term 6.58% $2,294 baseline
5/1 ARM, years 1–5 5.83% $2,119 –$175
ARM after reset, rate falls 1 point 4.83% $1,921 –$373
ARM after reset, rate unchanged 5.83% $2,119 –$175
ARM after reset, rate meets today’s fixed 6.58% $2,274 –$21
ARM after reset, first-adjustment cap hit 7.83% $2,543 +$248
ARM after reset, lifetime cap hit 10.83% $3,235 +$941
Thirty-year term throughout. The 30-year fixed rate is 6.58%, the Freddie Mac survey average for the week ending 23 July 2026. Freddie Mac’s weekly survey no longer publishes a 5/1 ARM average, so the introductory rate is shown as an illustrative 0.75-point discount rather than a quoted market rate; your own margin, index and caps will differ and are set out in your loan documents. Reset payments re-amortise the $334,284 balance remaining after 60 payments across the remaining 25 years. Caps follow the common 2/2/5 structure. Amortisation calculated by The Housing Signal.

Read the table as a distribution of outcomes rather than a recommendation. The introductory discount is real money — roughly $175 a month, or about $10,500 across the fixed period on this loan. What you are buying with it is exposure to whatever rates are in five years, capped but not eliminated. If the reset lands near today’s fixed rate, the ARM has cost you nothing and saved you the early years. If the first-adjustment cap is reached, the payment rises by roughly $250 a month against the fixed alternative and stays elevated for the remaining term unless you refinance. At the lifetime cap the payment is more than $900 a month higher.

Three points decide whether that trade is sensible for a given household. The first is time horizon: if you are confident you will sell or refinance inside five years, the exposure may never arrive, though confidence about a five-year horizon is worth less than most buyers assume. The second is whether the household could absorb the capped payment if it did arrive, because a refinance is not guaranteed to be available on acceptable terms in the year you need it. The third is the structure of the specific loan, meaning the index, the margin, the caps and the adjustment frequency, all of which vary by lender and all of which are set out in the loan documents rather than in the advertised rate.

An ARM is a defensible product for a borrower who understands the exposure and can carry the worst case. It is a poor product for a borrower who needs the introductory payment in order to qualify.

What Forecasters Are Saying, and What That Is Worth

The Mortgage Bankers Association, Fannie Mae, Freddie Mac and the National Association of Realtors all publish regular rate forecasts, and the consensus through 2026 has been for rates to stay within a relatively narrow band, with the direction contingent on the inflation path. Those forecasts are useful as a summary of what informed institutions currently expect. They are not useful as a plan.

The record supports the caution. Consensus forecasts materially underestimated the 2022 repricing, then repeatedly called for declines through 2023 and 2024 that arrived later and more shallowly than projected. This is not a criticism of the forecasters; rates depend on inflation, policy and global capital flows, and forecasting all three eighteen months out is not a solved problem. It is an argument for treating any published rate forecast as one scenario among several rather than as a schedule you can build a purchase around.

What This Means for Buyers

The practical conclusion is to stop trying to time the rate and to control what is controllable. Credit score, debt-to-income ratio, down payment size and documentation readiness all affect the rate you are personally offered, and the difference between credit tiers is frequently larger than the movement people are waiting for.

Collect more than one quote on the same day. The spread between lenders for an identical borrower routinely reaches half a point, which is the same $120 a month discussed at the top of this article, available without any change in market conditions. Compare the annual percentage rate and the lender fee schedule, not the headline rate alone. Our guide to reading a mortgage calculator sets out what the standard output leaves out, and the mortgage payment calculator will run your own figures.

Understand your rate lock. A lock of 30 to 60 days protects you between contract and closing, extensions usually cost money, and some lenders offer a float-down that lets you capture a decline after locking. Ask what the float-down costs and what trigger it requires, because an unused option is not free.

Finally, weigh the cost of waiting honestly. Lower rates increase what every other buyer can bid, and demand released into constrained inventory has historically pushed prices up faster than payments come down. A cheaper rate on a more expensive house bought against more competition is not obviously a better outcome.

What This Means for Current Homeowners

If you borrowed at a materially higher rate than today’s, the standard threshold is a reduction of roughly 0.5 to 0.75 points combined with a hold period long enough to recover closing costs. The break-even calculation is closing costs divided by monthly saving; if you expect to stay past that point, the refinance generally pays. Our refinance calculator runs that break-even against your actual balance, rate and closing costs.

Two adjustments to that rule are worth making. A new 30-year term restarts amortisation, so a refinance that clears the break-even test can still raise total interest paid over the life of the debt — compare remaining interest, not just the monthly payment. And rolling closing costs into the balance finances them at the new rate rather than avoiding them, which lengthens the real recovery period.

For the roughly one in five outstanding mortgages originated at sub-4% rates in 2020 and 2021, the calculation does not arise. Those loans are worth keeping, and the only reasons to disturb them are structural: accessing equity, removing mortgage insurance, or changing the term deliberately.

The Bottom Line

Rates through the remainder of 2026 will follow the inflation data, the labor market and the bond market’s read on both. They have drifted up by about half a point since February, they sit near the middle of their post-2019 range, and no one publishing a forecast knows where they finish the year. The decision that survives all of those unknowns is the same one it always was: buy a house you can afford at the payment you can actually get, and treat the rate as a variable to manage rather than an event to wait for.

Related reading: for how these rates are playing out in prices and inventory, see our 2026 housing market update. If you are weighing whether to buy now or keep renting, our rent versus buy analysis sets out the break-even maths. You can also track mortgage-rate movement over time with our plain-English updates.

Written by Mouhssine Ezzidi — Independent researcher · Founder & Principal Editor, The Housing Signal

With a background in financial data analysis, Mouhssine focuses on breaking down complex housing market trends into transparent, primary-source calculations. He builds The Housing Signal’s calculators and writes its mortgage analysis, working from Freddie Mac’s PMMS, CFPB guidance, and Federal Reserve data. He holds no mortgage license and sells no financial products. The Housing Signal is an independent publisher, not a mortgage broker or lender, and does not accept payment to rank lenders or steer readers toward specific products. Editorial standards

Disclaimer: The Housing Signal is an independent educational publisher. We are not a mortgage broker, lender, or licensed financial advisor, and nothing here is personalized financial advice. Rates, programs, costs and market conditions change frequently, and national figures may not reflect your local market or your own circumstances. Figures cited are accurate as of the dates given and will change. Consult a licensed professional before making a decision.

Advertisement (728x90)

Related Articles