A 6.85% average rate on 30-year fixed mortgages has pushed a growing share of US borrowers towards adjustable-rate mortgages, even as total mortgage applications decline. The shift points to a housing market where affordability pressures are changing not only how many people seek loans, but also the kinds of risks they are willing to accept to enter or remain in the market.
The Mortgage Bankers Association said total mortgage applications fell 2.7% in the latest week covered by its seasonally adjusted index. Applications to refinance existing home loans dropped 6%, while home-purchase applications declined by a more modest 0.2%. At the same time, ARMs accounted for 8.5% of all mortgage applications, up from 8% the previous week and the highest share since June.
The immediate trigger was a rise in the average rate for a 30-year fixed mortgage. The rate increased to 6.85% from 6.79% a week earlier, according to the MBA. The latest figure applied to conforming loan balances of $832,750 or less and assumed a 20% down payment. Mortgage points, including the origination fee, also rose to an average of 0.67 from 0.65.
The figures matter because the 30-year fixed mortgage has traditionally offered borrowers payment certainty over a long period. When its cost rises, households face higher monthly obligations or must reduce the amount they can borrow. Some borrowers may postpone a purchase, while others may look for loan structures with a lower initial payment. The latest MBA data suggests that both responses are occurring, although the decline in purchase applications was smaller than the fall in refinancing activity.
The growing interest in ARMs reflects the gap between initial borrowing costs. The average rate on a five-year ARM fell to 5.82% from 5.94% during the same week. That was more than one percentage point below the 30-year fixed rate. For a borrower focused on the initial payment, the difference can make an ARM appear more accessible than a conventional fixed-rate loan.
That lower starting rate is not a permanent guarantee. An ARM can keep its initial rate fixed for a specified period before the rate changes. Some loans offer an initial fixed period of as long as 10 years, but the rate can still change afterwards according to the terms of the loan. Borrowers therefore exchange some immediate payment relief for exposure to future rate adjustments.
The current rise in ARM demand is notable when compared with the early pandemic period. When mortgage rates fell to record lows during the first years of the COVID-19 pandemic, ARMs represented about 3% of mortgage applications. Their 8.5% share today remains a minority of the market, but it is substantially higher than that earlier level. The comparison shows how a different interest-rate environment can alter borrower preferences even when the underlying housing need remains.
The data also separates two parts of the housing market that respond differently to borrowing costs. Refinance applications were 25% lower than during the same week a year earlier and reached their slowest level since May 2025. Refinancing becomes less attractive when existing homeowners cannot secure a meaningfully lower rate than the one on their current loan. The 6% weekly decline therefore reflects a direct reduction in the number of loans for which refinancing makes financial sense.
Purchase applications were more resilient. Although they declined 0.2% from the previous week, they remained 4% above the level recorded during the same week a year earlier. This suggests that demand to buy homes has not disappeared. Instead, prospective buyers appear to be operating in a market where higher financing costs constrain their choices. The supplied data does not establish whether those buyers are reducing budgets, delaying purchases, choosing smaller homes or turning to different loan products.
The distinction is important for understanding housing demand. A fall in total applications does not necessarily mean that every part of the market is contracting at the same rate. Refinancing activity is more immediately tied to the relationship between current and existing loan rates, while purchase activity also depends on household formation, income, available homes and the decision to move. In the latest figures, refinancing absorbed the sharper decline, while purchase applications remained above their year-earlier level.
The report attributes the rise in mortgage rates to investor concerns about inflation and the federal budget deficit. Joel Kan, the MBA’s vice president and deputy chief economist, said those concerns were among the main reasons rates moved higher. This places the housing-market change within a broader financial environment: mortgage pricing is affected not only by local home prices or borrower demand, but also by how investors assess inflation and public finances.
That connection makes the housing market sensitive to information beyond the property sector. Mortgage rates were reported to be unchanged at the start of the following week in a separate survey by Mortgage News Daily, while investors awaited new monthly inflation data. The report said the inflation release could influence rates in either direction depending on whether the figure was higher or lower than expected. The supplied material does not establish the subsequent market outcome, but it shows why borrowers and lenders were monitoring the release.
For housing institutions and market participants, the present pattern raises a question about the distribution of risk. A lower initial ARM rate can help a borrower qualify for a loan or reduce early payments, but the possibility of later increases transfers some uncertainty into the future. The current data shows greater use of ARMs, but it does not show how borrowers evaluated the reset risk, how long they expect to hold their homes or whether they plan to refinance before the initial period ends.
Those unanswered questions limit what can be concluded from the application figures alone. The MBA data records loan demand and product selection, not household financial resilience. It does not indicate the income profile of ARM borrowers, the proportion choosing five-year products compared with longer initial fixed periods, or the number of applications that ultimately become completed loans. It also does not establish whether the increase in ARM demand will persist.
The figures nevertheless provide a clear account of the immediate direction of the market. Fixed-rate borrowing became more expensive, total mortgage applications declined and the share of applications for adjustable-rate products increased. The change was most pronounced in refinancing, while purchase applications remained comparatively stable and above their level from a year earlier.
The broader urban and housing question is how long households can continue to absorb higher financing costs before demand changes more sharply. The report notes that housing inventory has increased in many markets, but also says higher mortgage rates continue to weigh on people seeking to buy homes. More available homes do not automatically translate into stronger transactions if financing costs reduce the number of households able or willing to borrow.
What the evidence confirms is a market under pressure from the cost of credit rather than a uniform collapse in housing demand. The 6.85% 30-year rate, the 2.7% weekly decline in total applications and the rise in ARM share together show borrowers adjusting to a more expensive lending environment. What remains uncertain is whether the adjustment will remain a temporary response to rate movements or become a more lasting shift towards loans with greater future-rate exposure. Upcoming inflation data, subsequent mortgage-rate readings and later application figures will determine how that transition develops.
