Demand stalled in March for new-home purchases: MBA

Mortgage applications for new homes stalled in March as buyers felt the brunt of sustained home price growth.

Compared to February 2024, mortgage applications for new-home purchases rose 1% in March. Applications were up 6.2% from March 2023, according to the Mortgage Bankers Association’s (MBA) Builder Application Survey.

According to MBA estimates, new single-family home sales were at a seasonally adjusted annual rate of 615,000 units in March, the slowest annualized pace in four months, and a decrease of 10.7% from the February rate of 689,000 units.  

“March is typically a month when new home purchases see a seasonal boost, but this year March applications for new home purchases saw less than a one percent increase over the prior month on an unadjusted basis,” Joel Kan, MBA’s vice president and deputy chief economist, said in a statement. 

“Applications were still ahead of last year’s pace, but at 6%, the annual growth rate was the slowest since September 2023. Homebuyers remain adversely impacted by strong home-price growth and mortgage rates hovering around 7%.

“The FHA share of applications did increase in March, exceeding 26%, compared to a 24% average for the prior 12 months. A higher FHA share can be a sign of more first-time buyer activity, but that segment of buyers is also more sensitive to affordability challenges.”

In February, conventional loans accounted for 63% of applications for new homes. Federal Housing Administration (FHA) loans accounted for 26.4% of applications, U.S. Department of Veteran Affairs (VA) loans took a 10.4% share, and U.S. Department of Agriculture (USDA) loans accounted for 0.3%. 

The average loan size for new homes decreased from $405,719 in February to $405,400 in March.

New construction of single-family homes receded in March, and multifamily housing starts also declined. Additionally, homebuilder confidence stagnated in April due to elevated mortgage rates, coupled with a stronger-than-expected inflation reading.