New-home sales rose a very strong 20.7% in March, and this was off February levels that were revised upward substantially. Even with the revisions, February new-home sales are still somewhat depressed. Nevertheless, the March gain from that level is impressive.
A month ago, in our commentary concerning housing and other data, we were skeptical that February Texas blizzards were much responsible for soft data then. However, the strong March gains in most indicators showed that blizzards were indeed responsible for at least some of the February softness.
With respect to homebuilding and new-home sales, the declines occurred in most regions, not just the South. What’s more, this was consistent with the patterns of preceding months, where new-home sales had stalled since last summer, and housing starts gradually established a peak in response to peaking sales.
As strong as today’s new-home sales and last week’s housing starts data were, we still would assert that homebuilding is peaking. Averaging the elevated March sales levels with February’s blizzard-depressed rate, sales over the last two months are right in line with the prevailing pace in place since July. Single-family starts are behaving similarly, though with an expected lag.
Furthermore, adjusted for owner-builds, recent levels of housing starts have been more than consistent with recent levels of new-home sales. (The scales in the chart are staggered to account for owner-builds, which show up in the starts data, but not in new-home sales.) If new-home sales don’t rise further from here, homebuilders will not be able to raise starts levels further without adding to unsold inventories of homes.
Meanwhile, the regional patterns of new-home sales tell a similar story. Sales in the West have been declining since October, while sales in the Northeast and Midwest have been toppy. The bulk of March sales gains were in the South, which by itself is fine in that the South accounts for more than half of US homebuilding. However, it was the South that suffered the effects of the February blizzards, so the outsized (40.2%) March sales gains there are mostly rebound from the blizzard.
We should add that we are not claiming any actual weakness in homebuilding. Even if new-home sales and housing starts are indeed peaking, they are doing so at strong levels, relative to the experience of the last 15 years. However, homebuilding contributes to GDP growth by rising, not by being merely at a strong level. If, indeed, homebuilding and manufacturing activity are topping out after strong late-2020 rebounds, it will be up to service sectors to drive strong GDP growth over the rest of this year, and these sectors are still shut down in much of the country.